40 Burst results for "Paul Sweeney"
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-22-2023 12:00
"With Bloomberg, you get the story behind the story, the story behind the global birth rate, behind your EV battery's environmental impact, behind sand, yeah, sand. You get context, and context changes everything. Go to Bloomberg .com to get context. I tell you, there are leagues of investors out there. There's companies out there. Yeah, for sure. Yeah. I mean, I think this is going to flame out very, very shortly. I think of racquetball back in the 80s. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This is Bloomberg Markets with Paul Sweeney and Matt Miller. We got a lot of green on the screen here, but the volume is light. We constantly underestimate the strength of the U .S. consumer. This is a market that's much more optimistic or bullish than maybe the central bankers are. Breaking market news and insight from Bloomberg experts. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. All right, coming up in this hour, we're going to check in with Matt Hendrickson. He's the senior equity analyst covering medical technology at Bloomberg Intelligence. We're going to talk about a company, Boston Scientific. They recently had their investor day and they got a lot of new products coming out, so we'll get the latest there. Plus, it looks like Microsoft is actually going to be able to buy this Activision thing. They're getting some love from the U .K. regulators. We'll check in with Jennifer Rhee. She's a senior antitrust analyst at Bloomberg Intelligence. Then we're going to talk about drones and the use in the military because it is certainly front page news there, especially in Ukraine. Pretty extensive use of drones. All right, let's kick this whole thing off with Charlie Peller.
Bloomberg Markets
Fresh update on "paul sweeney" discussed on Bloomberg Markets
"When you get your news from Bloomberg, you don't just get the story. You get the story behind the story. How your EV's battery may not be as green as it seems. Why a decrease in global birth rates could send countries scrambling to increase immigration. You get context. Context changes how you see things. How you change things. Because context changes everything. Go to bloomberg This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. Matt Miller, Paul Sweeney Live here in our Bloomberg Interactive Broker Studio. Looking at these markets here. Up about a half a percent on the S &P 500. Not too bad. NASDAQ doing a little bit better here. Up nine tenths of one percent. So we're at or near the intraday highs on
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-21-2023 12:00
"2013 and now they're at 80 % Wow in terms of all new vehicle sales are full EVs But in the u .s.. Something seems to be holding us back right now about 7 % of new car sales or EVs But we're waiting to get to that tipping point where all of a sudden. We're off to the races We still have not hit it nice the Norwegians leaving the way and EVs how cool is that? Yeah, I just got the regular old gasp happy as a clam. We'll see how it plays out. This is Bloomberg Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg radio This is Bloomberg markets with Paul Sweeney at met Miller We got a lot of green on the screen here, but the volume is light We constantly underestimate the strength of the u .s.. Consumer This is a market that's much more optimistic or bullish than maybe its central bankers are breaking market news and insight from Bloomberg experts There's still some concern out there in the market that there is groups where things to deteriorate a little bit more than what they're indicating As small and medium -sized businesses struggle they don't present as much competition the supply chain has still got dislocations Globally and here in the u .s..
Bloomberg Markets
Fresh update on "paul sweeney" discussed on Bloomberg Markets
"Of 1 percent. I'm Charlie Pellett and that is a Bloomberg Business Flash. Charlie Pellett thank you so much we appreciate that. Matt Miller Paul Sweeney live here in the Bloomberg Interactive Brokers studio and we're also streaming live on YouTube. You know I was just scrolling through some social I saw Dan Ives you know he's at the perma bull on Tesla and his calls simply here you know what's bad for the big three is good for Tesla. You got to believe this is good for Tesla. I mean I was just looking at some data showing how many days supply car dealers have. So right now the average car sits on a lot for 37 days yep but you know that's well off the industry
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-19-2023 12:00
"With Bloomberg, you get the story behind the story, the story behind the global birth rate, behind your EV battery's environmental impact, behind sand, yeah, sand, you get context. And context changes everything. Go to Bloomberg .com to get context. Buy that? Yeah, maybe they planted trees. Yeah, okay. And then they flew here. S &P 500 off six -tenths of one percent, NASDAQ off seven -tenths of one percent, 10 -year treasury 4 .33 percent. This is Bloomberg. There's no concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. All right, coming up in this hour, we are waiting for Ukrainian President Volodymyr Zelensky to speak at the United Nations. When he does, we will bring that to you live. We're also going to check in with Siobhan Sutton, senior investment director, sustainable impact investing research at Cambridge Associates. Talk all about that, the climate, because it is climate week. It is climate week. So we're down with that. And the Instacart IPO, eventually it's going to get priced. We're going to check in with Mandeep Singh. He's a technology analyst at Bloomberg Intelligence. Get his thoughts on this company as he compares it to some of its peers as we wait for the trading of that one. But right now, let's kick this hour off with Charlie.
Bloomberg Markets
Fresh update on "paul sweeney" discussed on Bloomberg Markets
"Breaks my heart and it's something that Feeding America is working to change. Each year, the Feeding America network of food banks rescues billions of pounds of good food that would have gone to waste and gives it to families in need. To help, visit FeedingAmerica .org. Brought to you by Feeding America Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This is Bloomberg Markets with Paul Sweeney and Matt Miller. We got a lot of green on the screen here, but the volume is light. We underestimate constantly the strength of the U .S. consumer. This is a market that's much more optimistic or bullish than maybe central bankers are. Breaking market news and insight from Bloomberg experts. There's still some concern out there in the market that there is room for things to vary a little bit more than what they're indicating. As small and medium -sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Earth Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. Good Friday morning from the Bloomberg Interactive Brokers studio in New York City. It is Friday. We made it. Coming up, we're gonna take a simulcast of UAW's Sean Faines. He has some Facebook live comments coming up in just moments. When he does that, we will bring that to you. We're gonna check in with Ed
Bloomberg Radio New York - Recording Feed
Monitor Show 23:00 09-18-2023 23:00
"Investment Advisors. Switch to interactive brokers for lowest cost global trading and turnkey custody solutions. No ticket charges and no conflicts of your interests at ibkr .com slash ria. And 130 Industries. And remember, you can access Bloomberg Intelligence through BI Go and the Terminal. I'm Alex Steele. And I'm Paul Sweeney. Stay with us. Today's top stories and global business headlines are coming up right now. Broadcasting 24 hours a day at Bloomberg dot com and the Bloomberg Business Act. This is Bloomberg Radio. Talks aimed at ending a united autoworker strike are underway for a second straight day with no breakthroughs reported. Appearing on MSNBC, UAW President Sean Fain said that progress has been slow as the two sides continue to meet. Nearly 13000 autoworkers remain on strike, demanding better pay and pension benefits. House Speaker Kevin McCarthy says President Biden will get what he wants if members of the GOP caucus remove him from the speakership. If you did a motion to vacate, you would have to a handful of Republicans work with Adam Schiff, Eric Swalwell, Ehan Omar to remove the speaker. And it would be exactly what the president wants. During an interview on Fox News Sunday Morning Futures, the California Republican said ousting him would shut down the house and stop the impeachment inquiry against President Biden. McCarthy also brought up what it took for him to secure the speaker's gavel in January. He was elected after a historic 15th ballot. McCarthy's comments come after Florida Republican Matt Gaetz vowed to bring forward a motion to remove the speaker unless he meets conservative demands on spending cuts and policy reforms. Los Angeles County Sheriff's Deputy Ryan Klinkenbroomer was killed in an ambush -style attack Saturday night near the Palmdale Sheriff's Station. Sheriff Robert Luna says they're searching for answers.
Bloomberg Surveillance
Fresh update on "paul sweeney" discussed on Bloomberg Surveillance
"Again treasury is a slightly lower after touching those multi -year highs S &P 500 ten points higher that's up about a quarter of a percent the Dow Jones industrial average 20 points lower right now and the NASDAQ 100 down up 80 points that's up half a percent we check the markets for you all day long right here on Bloomberg radio I'm John Tucker and that is your Bloomberg Business Flash Alex and Paul. All right John Tucker thank you so much we appreciate it Alex Steele sitting in for Tom King Paul Sweeney here in the Bloomberg Interactive broker studio I think they'll be back the surveillance folks back in New York next week yeah I mean if they let them on the plane if if they they let him on the plane and I'm not sure if they're seeing any soccer matches this weekend or what the plan is. Whoa whoa whoa you cannot say I don't know you have to say football right football matches absolutely footy so anyway we'll have Tom back here on
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-15-2023 12:00
"With Bloomberg, you get the story behind the story, the story behind the global birth rate, behind your EV batteries, environmental impact behind sand. Yeah, sand. You get context and context changes everything. Go to Bloomberg .com to get context. Off about one point three percent will stay on top of that. We're still waiting for President Biden to make some comments. So when and if that occurs, we will bring that to you. We're going to have more coming up. This is Bloomberg. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. All right, coming up in this hour, we're going to check in with Ed Corey. He's a radio anchor and reporter at Bloomberg Radio. He's at the Detroit Motor Show. Oh, man, they must be talking in the corners there. We'll get the latest on what's happened in there. Then Barry Ritholtz, Ritholtz Wealth Management, we'll talk to him and see what he wants. What's on top of his mind, he is the host of Masters in Business, a big, big popular podcast. And then Todd Harrison, founding partner and CIO at CB1 Capital Management, discusses marijuana investing and the outlook for the cannabis industry. That's going to be cool. Todd Harrison is pretty legendary on Wall Street. He started Minionville.
Bloomberg Surveillance
Fresh update on "paul sweeney" discussed on Bloomberg Surveillance
"Bear because he's got more to say than just... Only You can prevent wildfires. Like if you're outside enjoying a barbecue. Don't let a hamburger distract you from fire safety. sure Make you aren't dumping your hot coals or ashes onto the ground because that could start a wildfire. So take wildfire seriously and let's save the world one day at a time. Go to SmokeyBear .com to learn more about wildfire prevention. Brought to you by the U .S. Forest Service, your state forester and the Ad Council. The data and analysis that give you an edge. Bloomberg Intelligence's ten companies to keep an eye on. The Bloomberg Intelligence podcast with Alex Steele and Paul Sweeney. Economic uncertainty remains a headwind for global bank stocks. In -depth research on 2 ,000 companies across 130 industries. We're gonna take a look at global life science Let's turn now to climate
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-14-2023 12:00
"Why do tacos get their own day of the week? Is it because Mondays are so rough we need a Tuesday filled with beefy tortillas shared with good friends? If so, why don't we have Wellington Wednesdays stroganoff Saturdays and heck, meatball Mondays? Then Mondays would just be another reason to enjoy our favorite beef with our favorite people. Together we bring more. Beef. It's What's For Dinner, funded by beef farmers and ranchers. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This is Bloomberg Markets with Paul Sweeney and Matt Miller. We got a lot of green on the screen here but the volume is light. We constantly underestimate the strength of the US consumer. This is a market that's much more optimistic or bullish than maybe the central bankers are. Breaking market news and insight from Bloomberg experts. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle they don't present as much competition. The supply chain has still got dislocations globally and here in the US. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. Alright coming up we're gonna check in with Mari Shore she's senior equity analyst with Columbia Thread Needle. We're gonna break down some of those retail sales numbers that we got today. They look pretty solid to me. On the VC outlook Robbie Peters co -founder Sempervirens Venture Capital discusses his firm and the outlook for the VC space and then Roberta Goss senior managing director and head of the bank loan and CLO platform at Predium will join us to talk about lending and the outlook for financing in the US. Right now we're gonna kick things off with Mr. Charlie Paul. Thank you very much Paul. Lots of moving parts in today's trading. We're keeping an eye out for arm holdings.
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-13-2023 12:00
"With Bloomberg, you get the story behind the story, the story behind the global birth rate, behind your EV battery's environmental impact, behind sand, yeah, sand, you get context. And context changes everything. Go to Bloomberg .com to get context. This is Bloomberg. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This Bloomberg is Markets with Paul Sweeney and Matt Miller. We got a lot of green on the screen here, but the volume is light. We constantly underestimate the strength of the U .S. consumer. This is a market that's much more optimistic or bullish than maybe the central bankers are. Breaking market news and insight from Bloomberg experts. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. Ay, ay, ay, very disturbing what's happening here in our Bloomberg studios, but we will soldier on. Coming up, we're going to check in with Jeffrey Cleveland. He's the chief economist at Payden & Riegel. Get his thoughts on the inflation print that we just saw. Brian King, CEO at Lotus Markets, will talk markets and ETFs plus tech talk. Actually, still, senior vice president and creative for digital media lead at Adobe. We'll discuss the company's generative AI initiative. So why not AI? More AI. Let's do it. But right now, let's kick things off with Charlie Pellett. All right. Thank you very much, Paul Sweeney. The Dow, the S &P, NoStack, they're all advancing right now.
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-12-2023 12:00
"With Bloomberg, you get the story behind the story. The story behind the global birth rate. Behind your EV batteries environmental impact. Behind sand. Yeah, sand. You get context. And context changes everything. Go to Bloomberg .com to get context. 24 that's according to ESPN. So I don't know somebody's on the hook for that 75 large. Yeah, but anyway, just a terrible, terrible story there. You feel terrible for him and for Jets nation as well. All right, we're gonna have more coming up. This is Bloomberg. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This Bloomberg is Markets with Paul Sweeney and Matt Miller. We got a lot of green on the screen here, but the volume is light. We constantly underestimate the strength of the U .S. consumer. This is a market that's much more optimistic or bullish than maybe the central bankers are. Breaking market news and insight from Bloomberg experts. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. All right. Coming up in this hour, we're going to talk to cruising business. The CEO of Viking Cruise is going to join us via Zoom, so that'll be very cool. We're going to get a preview of what we're going to hear out of Cupertino from Anurag Rana, Bloomberg Intelligence, then Adam Koons, Portfolio Manager, Winthrop Capital Management. I am going to talk to them, not we. I am going to talk to him about the markets coming up and what he's expecting with the CPI stuff. So that'll be a fascinating discussion. That's because I split early. That's because you split early to go do your TV thing. That's okay. Charlie Pellet though, he stays with us. He's got his news for us.
Bloomberg Radio New York - Recording Feed
Monitor Show 23:00 09-11-2023 23:00
"Interactive brokers' clients earn up to USD 4 .83 % on their uninvested, instantly available cash balances. Rates subject to change. Visit ibkr .com slash interest rates to learn more. And 130 industries. And remember, you can access Bloomberg Intelligence through BI Go on the terminal. I'm Alex Steele. And I'm Paul Sweeney. Stay with us. Today's top stories and global business headlines are coming up right now. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. The 22nd anniversary of the 9 -11 terror attacks is tomorrow, with many ceremonies planned around the country. Some family members who lost loved ones on September 11, 2001, are still left searching for answers over two decades later. Brett Eagleson's father was killed in the World Trade Center and doesn't feel victims' family members have all the details of who was responsible for the attacks. It's been 22 years and it's shameful that we have yet to have an administration or a president that has had the courage to hold the kingdom of Saudi Arabia accountable for what they did. Saudi Arabia denies involvement in 9 -11. Meanwhile, family members are also frustrated that the confessed mastermind of the 2001 terror attacks and four other defendants have not yet had a trial. Vice President Kamala Harris says the attacks on her by GOP presidential candidates are nothing new. Speaking on CBS's Face the Nation, Harris responded to concerns of President Biden's age and claims that she would be worse than Biden. The vice president said she's listened to attacks against her throughout her career, going back to her days as a district attorney in San Francisco. Harris said Biden would be fine to serve a full second term as president, but added she's prepared to step in if necessary. Pennsylvania State Police are expanding their search for an escaped killer after a new sighting of a woman was reported.
Bloomberg Radio New York - Recording Feed
Monitor Show 13:00 09-09-2023 13:00
"At Morgan Stanley, old school hard work meets bold new thinking to help you see untapped possibilities and relentlessly work with you to make them real. To learn more, visit Morgan Stanley dot com slash y us. Investing involves risk. Morgan Stanley Smith Barney LLC. On 2000 companies and 130 industries. And remember, you can access Bloomberg Intelligence through BI Go in the terminal. I'm Alex Steele. And I'm Paul Sweeney. Stay with us. Today's top stories and global business headlines are coming up right now. Broadcasting 24 hours a day at Bloomberg dot com and the Bloomberg Business Act. This is Bloomberg Radio. The death toll now topped 1000 after a powerful earthquake struck Morocco. Hundreds more were injured by the rare six point eight magnitude quake that struck late Friday night. Search and rescue operations are underway amid the rubble of ancient buildings in the historic city of Marrakesh. President Biden is among the world leaders pledging support. Biden, who is attending the G20 summit in India, issued a statement saying he is deeply saddened by the loss of life and devastation. He also vowed to work expeditiously to ensure American citizens in Morocco are safe. A federal judge is denying a request from former White House chief of staff Mark Meadows to move the Georgia election interference case against him to federal court. Brian Shook has more. Meadows argued on the grounds that he was a federal officer when the alleged election crimes took place. The move was seen as a better path for Meadows to perhaps get the charges dismissed. Last month, Meadows, former President Trump and 17 others were charged with felony racketeering and conspiracy counts in the 2020 presidential election in Georgia. The most intense Atlantic hurricane so far this season is on a track that would keep it away from Florida, but it could still bring dangerous conditions.
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-08-2023 12:00
"It says it was developed by Saul Kurtzner, who is apparently a storied South African hotelier, and has hosted New Year's Eve events headlined by performers including Fergie and Sting. Nice. I mean, the water park looks awesome. It's got a casino. It's got a gajillion hotels. So I would think people can continue going to this place no matter what, because it looks like a one -of -a -kind property in the Bahamas. What could go wrong in the Bahamas? This is Bloomberg. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This is Bloomberg Markets. With Paul Sweeney and Matt Miller. We got a lot of green on the screen here, but the volume is light. We constantly underestimate the strength of the US consumer. This is a market that's much more optimistic or bullish than maybe central bankers are. Breaking market news and insight from Bloomberg experts. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the US. This is Bloomberg Markets. With Paul Sweeney and Matt Miller. On Bloomberg Radio. All right, coming up in this hour, we're going to check in with Catherine Lim, senior analyst. She covers consumer and technology with Bloomberg Intelligence, because I want to talk to her about this company that's coming public, AMER Sports. They make Wilson tennis rackets and Solomon ski boots. Those are my favorite ski boots. Like that stuff. And plus, we're going to talk about Matt Miller. What is Matt Miller driving these days? You know who we got? Michael Dean. He covers all the Euro autos. Yeah. So he's like one of your favorite guys. We'll get him in there. We'll talk about Matt, what he's driving, because he's always driving something. And he's a total car guy himself. Total car guy. He loves cars. Yeah, he loves cars. Michael, Kevin.
Bloomberg Radio New York - Recording Feed
Monitor Show 15:00 09-07-2023 15:00
"So this is a bipartisan love affair with fast food that we've identified. Yeah. I mean, who doesn't like french fries? Absolutely. I'm just not sure about mixing it with cherry pie. Yeah. You know they put chocolate on them in Japan? I did not know that. Yeah. Every country has its own little McDonald's quirk. Yeah. Would you do the chocolate or the mayo in Europe? Mayo. Okay. Hold that thought. Kaylee, thank you. Bloomberg Business Week starts right now. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This is Bloomberg Business Week. Insight from the reporters and editors who bring you America's most trusted business magazine. Plus, global business, finance, and tech news as it happens. Bloomberg Business Week with Caro Masur and Tim Stenebeck on Bloomberg Radio. All right. Good Thursday afternoon from the Bloomberg Interactive Brokers Studio. Jess Menton and Paul Sweeney sitting in for Tim Stenebeck and Caro Masur. Who's who? I'm not sure who's sitting in for who. But Jess and I are here, so we're in good shape to go for the next three hours. There's a lot coming up, Jess. There's a lot. We've kind of got the markets are kind of mixed here on this Thursday afternoon, but a lot coming up in the next few hours. The highlight for me is... Oh, I have a guess which one it is. Jeffrey Katzenberg, he's a co -founder of DreamWorks Animation. He's just a major, major player in Hollywood. Has been for decades. He's going to join us. He just recently joined the board of a cyber security firm. Aura is the name of that. So we'll talk about that. We'll talk about... I'm going to ask him about the writer's strike. And then maybe this whole streaming thing. It's just upended the whole Hollywood model. Yeah, exactly. He was at Disney. We ran their studios successfully for many years. So that's just one of the great, great guests we got coming up. And then of course...
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-07-2023 12:00
"The United States Border Patrol has exciting and rewarding career opportunities with the nation's largest law enforcement organization. Border Patrol agents enjoy great pay, outstanding federal benefits, and up to $20 ,000 in recruitment incentives for newly appointed agents. If you are looking for a way to serve something greater than yourself, consider the United States Border Patrol. Learn more online at cbp .gov slash careers slash USBP. That's cbp .gov slash careers slash USBP. Booze, this is Bloomberg. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. Matt Miller here at 731 Lex with Simone Foxman. Paul Sweeney is going to join us later on this afternoon on Bloomberg Businessweek. We were talking about the fact that the market feels like it's off, and the reason that it feels, I think, like it's down further than it is is because it's a shortened trading week, and we've had sell -offs every day this week, right? We were down Tuesday, Wednesday, and today, and on Monday obviously closed for Labor Day. And we know that September is supposed to be negative for the S &P 500 over the past 30 years, down 710.
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-06-2023 12:00
"Morgan Stanley and RBC which is Royal Bank of Canada. They are Enbridge's financial advisors on the deals while on the sell side Citigroup and Goldman Sachs represented Dominion so that's good to see. McGuire Woods a Richmond based law firm also advising Dominion which is a Richmond based company too so lots of good stuff there. M &A in the energy space coming at you. We're gonna have more. This is Bloomberg. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This is Bloomberg Markets with Paul Sweeney and Matt Miller. We got a lot of green on the screen here but the volume is light. We constantly underestimate the strength of the U .S. consumer. This is a market that's much more optimistic or bullish than maybe central bankers are. Breaking market news and insight from Bloomberg experts. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. Alright coming up in this hour we're going to check in with Brian Egger Bloomberg Intelligence. He covers all the casino stocks, cruise lines, all that kind of fun stuff. We want to get an update on what's happening in Vegas post -pandemic Vegas baby. Diana Scott lead of U .S. Human Capital Center at the conference board is going to join us to discuss its recent findings on hiring in corporate America. So lots coming up in this hour as well but first let's kick things off with Mr. Charlie Pellegrino. Hi thank you very much Paul Sweeney and here's what's going on to Dow the S &P Nez Stack all in the red with the S &P extending its drop below 4 ,500 right to the numbers with the S &P at.
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-05-2023 12:00
"I'm not a big sugar guy. I use the sugar substitutes, like the blue stuff and then the yellow stuff. I eat way too much sugar. I really, I'm gonna start, like, today. Today. Cutting back. Oh, you heard that here first, Matt Miller. Cutting back on sugar today. I came home from Spain last night. First thing I did, I ate four Italian ices when I got home. They don't have Italian ices in Spain? No. No Italian ices in Spain? They're delicious here. Go figure. Uh, S &P 500, off two -tenths of one percent. We're gonna have more coming up. This is Bloomberg Markets. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This is Bloomberg Markets with Paul Sweeney and Matt Miller. We got a lot of green on the screen here, but the volume is light. We constantly underestimate the strength of the U .S. consumer. This is a market that's much more optimistic or bullish than maybe central bankers are. Breaking market news and insight from Bloomberg experts. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. All right, coming up in this hour, we're gonna check in with Fernando Valle. He's the senior analyst with Bloomberg Intelligence covering the global oil space. We got Brent crude at over $90 a barrel. We need to figure out what's going on there. Fernando will have the answer. And as we mentioned earlier, we're gonna check in with Nakul Nagal. He is Qualcomm's senior VP and he does all the automotive business for Qualcomm. So we'll get a sense of what's going on there as the chips become a bigger part of the auto business. And then Katie Charleston, founder of Charleston Law. She's gonna join us to discuss legal concerns over AI as it relates to the...
Bloomberg Radio New York - Recording Feed
Monitor Show 13:00 09-01-2023 13:00
"In coordinates as far as some of the news that we've gotten out of some of these weight loss drugs but always great stuff uh Simone getting your perspective on all things this and always a pleasure being able to co -host with you as well. Oh thanks it's a pleasure being able to co -host with you too ahead of the Labor Day weekend. Right we do hear uh Matt Miller, Paul Sweeney we should be back next week so stay tuned for that but always a pleasure here but Jess Moon Foxman Sound On with Joe Mathew starts right now. We're totally different from four years that Biden has been there people are going to say I was doing better then than I am now. Bloomberg Sound On politics policy and perspective from DC's top name. So you got to work to get people back to work but not only that but to higher paying jobs. The Russian threat is being degraded and unfortunately it's being degraded at the cost of Ukrainian lives, blood, treasure. Bloomberg Sound On with Joe Mathew on Bloomberg Radio. Is this the soft landing or the peak? Welcome to the fastest show in politics on this jobs day. We'll review another apparent Goldilocks report the White House is celebrating with the acting Secretary of Labor Julie Hsu with us in a moment here and we'll also take a deep dive with Mark Zandi the chief economist at Moody's Analytics who seems pretty upbeat about things so far today. President Biden as I mentioned celebrating the news we'll talk about that with our political panel. Bloomberg politics contributor democratic analyst Jeannie Shanzano joined today by Republican strategist Lisa Camuso Miller. She's back with us from Reset Public Affairs the former community.
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 09-01-2023 12:00
"I mean, eventually at some point there will be a recession, whether or not it's in the... And that will shift the dynamics there to where people will be coming to the office. Not saying that it is coming soon, but eventually in a business cycle it will happen and people will have to start coming up. I just want to put out the warning out there if you're near the link and don't pull over because I'm coming. I have to say that quickly, John Tucker unfortunately is leaving us, but it's been a pleasure. Simone Foxman's coming up next with us, so everybody stay tuned, but thank you for doing the last two hours. That was great. That was a lot of fun. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle, they don't present as much competition. The supply chain has still got dislocations globally and here in the U .S. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. Jasmine and Simone Foxman here in the Bloomberg Interactive Broker Studio. She's back. She was here a couple of hours ago, so she's up next with us for the next hour. Of course, as we know, Matt Miller, Paul Sweeney do both have the day off, so stick with us over our final hour for Bloomberg Markets. We have Jennifer Lee of Bloomberg Intelligence joining us to talk about the FTC settlement on the Amgen Horizon deal. Also, Eddie van der Walt from Bloomberg News. On latest with the jobs data and Martha Heller from Heller Search Associates on the outlook for tech hiring. But first, let's get a Business Flash with Charlie Pellet. Thank you very much. Yes.
Bloomberg Radio New York - Recording Feed
Monitor Show 12:00 08-31-2023 12:00
"Three billion dollars and then they upped that to four and a half billion dollars so the costs associated with them are pretty steep. And I wonder how this plays into ongoing discussions with the United Auto Workers unions with between General Motors, Ford, Stellantis. Yeah absolutely I mean there's a big big discussion point there about EVs and what it means for the workforce of course for these automakers so we'll have more reporting on that coming up. This is Bloomberg. Broadcasting 24 hours a day at Bloomberg .com and the Bloomberg Business Act. This is Bloomberg Radio. This is Bloomberg Markets with Paul Sweeney and Matt Miller. We got a lot of green on the screen here but the volume is light. We constantly underestimate the strength of the US consumer. This is a market that's much more optimistic or bullish than maybe its central bankers are. Breaking market news and insight from Bloomberg experts. There's still some concern out there in the market that there is room for things to deteriorate a little bit more than what they're indicating. As small and medium -sized businesses struggle they don't present as much competition. The supply chain has still got dislocations globally and here in the US. This is Bloomberg Markets with Paul Sweeney and Matt Miller on Bloomberg Radio. Alright coming up in this hour we're gonna talk to Molly Smith she's a US economy editor. We'll talk a little bit of economic news but I really want to get her thoughts on the US Open. She's the big tennis fanatic here in the building so we'll get the latest what's happening out at the US Open in Queens. And then we're gonna talk about the retailers. Brendan Case US retail reporter with Bloomberg News. We're talking about Dollar General's earnings miss and I'm gonna preview the Lululemon earnings coming after the close today. Then ETFs you can never talk about ETFs too much because that's where the money's going. Bill Davis managing partner and portfolio manager at Stance Capital. He's gonna join us to talk about ETF investing where the money is in fact going what strategies are.
VUX World
"paul sweeney" Discussed on VUX World
"So I think it's going to be, like, what I think is interesting is it has a different mindset, it has maybe a different design. It's definitely going to have a different approach to programming. Like, if you can, I don't know if you've played around with the programming aspects, but I actually took one dialogue engine in the market and I said, hey, could you write code to translate this from Google to Microsoft? I said, yeah, there you go. Like, what? Does that run? Yeah, it does. It's like, oh, my God. Once you get into it, I think it's very interesting. I think the idea of emergence is super interesting. I don't know if it's true or not, that it has emergent properties. That could be, like, that could be really interesting. I don't know how much of the model, like, concept information it's able to actually hold versus it's just, you know, statistics driving forward. Like, this is the next probable thing, this is the next probable thing. So I think these abstractions are going to end up being really important at the end of the day as well to get it to the next level. Yeah, absolutely. And the models will improve as well, you know, as they advance from simply being language in, language out, and they start to be able to call APIs, they start to be able to do other kinds of things. Inevitably, I can see a world where rather than having session-based context, which is if you refer to something in a previous line, it will understand what you're talking about. It has a very loose grasp of context. I can see a world where memory management becomes built in. At the moment, you've got companies within the ability to have memory, like Langchain and Voiceflow. I can imagine this becoming part of the model, part of the API. You can decide what you want to store in memory, you can decide what you want to feed into context, you can give it specific things that it can reason on. At the moment, it will reason within a degree, but it won't reason about something in your very specific collections because it doesn't know what that process is. Well, it could. It could. Again, it all depends on how these things are architected going forward and how much privacy you have and how much governance you have over the thing. So I can see that if you've trained a language model and say, hey, here's the process, don't deviate from the process. Okay. Here's language. Don't use language that deviates between minus 10-year-old and plus 15-year-old. All your language should be within that language understanding frame. Cool. And it just runs like that, or you give it little hot and cold measures, like deviate a little but not too much to generate difference and then feed that back in. Was one effective or not? So I agree. And I think what it feels like is I'm trying to figure out is it more like cloud or is it more like database? I'm trying to figure out which one of those it is, but it definitely feels like infrastructure. And then there'll be various tools sitting on top of it to monitor it or keep it compliant or to keep it private, to keep it secure, to do other tasks with it, or even to memory manage, to cost control it. These things aren't cheap to run, but that cost will come straight down as well. So yeah, I think to round up our view on it is in financial services, they'll be very careful for a while. They'll be very careful about if you know anyone who sells into banks, they're super careful about what they put in their infrastructure. I think they will need really secure, really guaranteed, fully auditable, and the language models just aren't there from using it as a service, like just taking your IT department and pointing them at ChatGPT. So I think they'll need someone to help them with that. Making sure that all your data in the enterprise is synced in and out properly, I think that's a medium-term opportunity still, and that's why we're so big on the API element here. You've got to be able to get the data in and out properly. So I don't think NLM is going to do that anytime soon. It's capable of it, just like we said it was now, but I just don't think it's going to be there in the near to medium term. And I think that then just the services on the front end of it, like what does it actually, like what's your front end? How are people using it? Like what tools do they have? How's it making their jobs better? What's the assistant doing with somebody else to make a job easier? I think that's like we're still waiting for a really good interface for that. I have some ideas around how that might happen, but we haven't really truly gone dynamic in the web interactions yet. Like it's still fairly flat. So I think if there is an opening, if something is dynamic and generative, and there is some controls on it, I think the next generation of the front ends will be interesting as well. Yeah, and the same could be said for not just the web, but mobile apps and any other system basically. The conversation as a UI, there's huge opportunities there. Most of the time on most websites, most of the chatbots are just little, as you alluded to in the past, little widgets in the bottom corner. You click on them and you have a conversation, or they pop up at the relevant time and say, let's chat, whatever. Whereas, and I've seen some examples of people who've tried to make their whole website a chat interface, but it's the same thing. It's just a chat interface on a website. It's just like a whole screen is a chatbot. But then I've seen other prototypes where it gets pretty impressive, where like half of the screen is the chat widget. Half of the screen is the website, but it's not necessarily the website. It's just when you ask a question, it will extract some of the data from the website and present it visually, if it's an interface, and then widgets will occur in the chat window. This is why it all comes down to building your own interface, because nothing out of the box comes like this. For example, if you wanted to have a little toggle that says, how much can you afford to pay per month? And you can just toggle zero up to 500 pounds or whatever. Little interfaces like that where you're still part of the conversation, but you're adding a level of engagement to try and people can set their own thing, and then when they say that, something visual happens here, and the amount flashes up and it says, okay, you can pay 50 pounds a month. So you start to create more dynamic experiences. I agree, yeah. Integrating the website with the conversation. It's just that most people, because of the intent-based models and the deterministic flows being quite laborious to get right, I think people have been focusing there rather than how do we make this experience the best experience, and maybe that's where LLMs can afford us scope. Did you see that example of the car advertisements where you've got a screen and you've got the car and someone says, well, what do you want to see? And you go, well, how fast does it go? Good question. It goes from here to here in this kind of weather, there to there in that kind of weather. Would you like to see it go? And then boom, you're in the car and it drives off, right, and you see you're on a road and going, well, that feels fairly good. Would you like a different interior? Would you like this in a wood interior? Oh, yeah, that'd look lovely. There it is, wood interior. And so it's just generating an experience, and I think it's like where an LLM has your data, like your height, your weight, your width, right, your physical size, and that's stored as a measure in your intelligent assistant somewhere. Now I can put you in against that car. I can put you in that seat. I can show how you actually fit in this environment. And I know we've always been, the last 10, 15 years, we've been talking about this like depersonalization in sales, but if somebody gets it right, like if somebody gets this right, a convenience always wins. Like the most, like if there's one thing out of everything that's gone on the web, like convenience wins. So how can you like make it more convenient, more convenient, more convenient? That's still the winning formula as far as I can see.Definitely. Reducing friction, saving people time, making everyone's lives easier is still the end of the game. It's just a new technology that we can do that with. That's been the case with mobile and web and all that kind of stuff. But I think that, yeah, this is definitely an exciting phase that we're in. I'm infused by how much excitement is present generally in the world now because everyone's woken up to conversational interfaces, which is fantastic. And I think that, yeah, it's a bit like the Wild West at the minute with lots of stuff going on, lots of people trying stuff, lots of kind of hype and all that kind of stuff, lots of ideas. And yeah, I'm excited to be a part of it and excited to have you on the podcast to be a part of this. So thank you so much, Paul, for your time. We've gone over on time there, but thank you for being with us. It's been an immense conversation. I told you it wouldn't be short. And thank you all for tuning in, Sandesh, Sean, Miguel, Lauren, Brett, you know, quite a lot of people commenting there. Apologies we didn't manage to get to all of those comments, but maybe we can do after we've finished broadcasting, we can follow up with some of that kind of stuff. Thank you, Tyrio, for sponsoring. And if you are heading out to Washington, D.C. in September on the 5th through 7th, you can witness the VideoX stage at the Voice and AI event. We'll be doing it on the 6th, on the p.m. afternoon. We've got a whole bunch of good companies coming, sharing some really great content. So if you want to get a ticket for that, you can go to voiceand.ai to find out more about that conference in Washington, D.C. this September. Thank you all for tuning in. Thank you again, Paul. I knew it would be pleasurable, but it's been absolutely immense. Thank you so much. Pleasure to be here.
VUX World
"paul sweeney" Discussed on VUX World
"Oh, definitely. I absolutely agree. I think what's really interesting is that most of the time, technology on its own, one piece of technology on its own isn't the thing that makes the actual difference. It's when you couple that with other technologies. Like high-speed internet on its own wasn't the thing that made the difference. It was the fact that you had devices that you had in your palm of your hand and social media sites or web infrastructure, cloud-based infrastructure coupled with the internet. Then all of a sudden you can watch videos on HD on your phone. It's not necessarily the internet on its own that's the issue or the driving factor. It's the fact that now you've got mobile connectivity and you've also got GPS and satellite coordination joined together so now you get the Ubers and the Halos of the world. So it's when you can take those technologies and bring them all together and then you get something special. So large language models are just a singular technology. So what happens when you start doing things like that, connecting them, giving them the ability to build applications, connecting them to certain data sources, adding other technologies around them that then turn it into something far more powerful than it would be on its own. I was just back in 2000. I remember doing a startup and doing a cloud-based startup in 2000 and the hosting costs for the application came were 250 grand a year. So to do a startup and just pay the hosting costs was 250 grand. Now given you can spin up Amazon instances really cheaply now, that opens up a whole range of new startups that can think about what they might be able to do in the evening and that will give people access to their first, second and third failure and fail cheaply and then get on to the next thing. So I think it does fall into that category. It's also, as you said, things are happening so frequently and the fact that ChatGPT has managed to get this on everyone's radar. Alexa was trying, people started building skills, there was no money in it, it stopped being exciting so people drifted away, whereas now this is well and truly going through that hype cycle and there's lots of people that are involved in building applications and experiment with this kind of stuff. A lot of people are involved in the research to try and prove out new ways and new methodologies around getting more performance in certain areas around these models. So it is because there's so many people interested in it, there's so many people able to do so much stuff. So it's very much exploratory. This is where the whole exponential curve comes because everyone's building stuff and everyone's experimenting and trying stuff. It's just like, it's mad. It's really crazy how fast things are moving and how many people are involved in trying to get things off the ground. I was just, again, as I said, I'll never tell you a short story, Cain. I was sitting, I was giving, I was an invited speaker to a class of American students, college students, and I was doing my talk and we're just talking at the end of it about what they wanted to do and what their backgrounds were, et cetera. And two of the students there asked them what they were doing and they said, oh, we're drama students. I went, brilliant. And they went, well, what do we do in this world where an AI writes the script and an AI does that? I said, no, just think about it. And just today, if you were thinking about how you interact with this chat GBT, you have to assume a persona, you have to set up a dialogue, you have to give them a situation. And actually, if you have improv skills, if you're the kind of person who's used to improv, you have a very flexible approach to how this conversation might be run, how I might set it up, how I might kind of, you know what, I won't ask if this is a straightforward, what would an average customer ask for in a bakery? Like, I'll set up a scene, you know, set the scene, it's a bakery, now write a script about whatever, whatever, and I get some interesting insights about how people interact because I've had a very flexible approach to it. So I think we might call this like some sort of affordances, like we don't really understand the affordances of the language models yet, like we kind of think about it as prompting, but then it's also acting, it's also staging, it's also scene setting, it's also kind of play, and you might be able to get different insights from, say, write this scenario from the point of view of a biologist and then write the scenario from the point of view of an economist, now have the biologist and the economist argue about whose approach is correct, now do something else, now do something else. And so these are not kind of skills that you would have 12, you know, six months ago going, these are going to be useful skills for your middle to senior management job here of doing scenario development for your industry. And I did put in to ChatGPT, I said, you know, let's write a paper on what's important for conversational AI and let's do a paper on it. And I know the area pretty well. You know what, it was pretty accurate. Like, it was pretty accurate. Like, if you weren't an expert in that area, that would be good enough. It would be good enough.
VUX World
"paul sweeney" Discussed on VUX World
"I think that there's more to this technology than just I can drop documents in it and what happens. That's the equivalent of saying I can go online and see a web page. It's like that's where we are with regards, I think, where this can go. So I'm not at all doomsdayish about it. I'm just saying I think it's a fundamental thing as opposed to it's faster at reading documents and it'll tell me what it's about. I think that's the immediate sugar, empty calorie burst. So I think it's a much bigger shift than we maybe understand right now and if someone tells you they know how it's all going to play out, you don't know how this is playing out on a daily basis. I'm following some of those papers and going, I'm just trying to think of one that was there yesterday that was just complete opposite. Oh, it's actually in a different area. I just saw a paper yesterday on that actual social media may not be dividing people at all. It's just reflecting the divisions that are already there. Now, if you were walking around the world as a human being for the last five years, you just go, Facebook blew up the world. They did everything. They're the cause of all this. Actually, they're reflecting what was there already and now you're able to see it and maybe that creates another cycle, but I think that the papers are changing every day and it's maybe the first time that such a huge shift in technology or anything has happened and all the data has been published immediately. So all the papers like open source, they're all getting published. They're all available to be replicated really quickly and I think when you connect that kind of an idea of that data being out there with open science approaches and open source communities, I think you could see this just develop very, very, very quickly and we don't know exactly what happens next. We just don't, but I'm not going, it's all doom and gloom. I just don't think it is going to be all doom and gloom. I think it's going to be very exciting.
VUX World
"paul sweeney" Discussed on VUX World
"Most financial services companies do not want to hold that data. No. It's not relevant to the conversation. In fact, they would probably be breaching GDPR if they were to gather that data because you can only gather and store data that's related to the thing that you're kind of serving people on. But I do want to ask you one more question. There's a few questions in the chat, and apologies if we didn't get to it. I know that we're past time now, but I did have one conversation around the use case that you were talking about where everyone's going to have their own co-pilot and every organization will have their own LLM and all that sort of stuff. I can definitely see that happening. I can see some of the obstacles that need to be overcome to get to that point. One of them is data quality. We're going through an exercise at the moment of taking all of the podcast transcripts, including this one. We transcribe them all anyway, just so that we have them. We're documenting all of this. Every step of the way will be shared publicly so people can follow us on this kind of journey. We want to build a custom language model that is specific to Vuex World that has all of the podcast content in it. So then you're going to be able to ask any question that's being covered on any podcast or any article that we've published related to conversation design or conversational AI or conversational automation, and it will give you an answer and it will source the podcast episode, tell you who the guest was. The first issue we have is that speech-to-text engines don't always transcribe things effectively, especially if you have a northern accent like me. So the first problem we have is data cleanliness. Businesses have lots of data in PDFs, lots of things in spreadsheets, lots of things in PowerPoint presentations, lots of things in business systems and knowledge bases. Some of it's out of date, some of it's within date, but the quality might be not so good. So how important is it that the data is up-to-date, accurate and sorted? Because that's arguably a massive job in itself, and an ongoing job in and of itself is data management. I might be out of my realm of expertise here, but I'm kind of impressed with the coding. I think that one of the big things with the LLMs is the coding capability. And, for instance, if you're able to say... Let me just kind of step back again for a second. If you can tell an LLM, hey, this is an API, like here's a forward slash, you know, vux.com forward slash API document. Okay, that's your published API for accessing your podcast data. And I can tell the LLM, build an interface to a VUX world using the VUX world data at this location, the API, and it can code up an interface. And then I can say, I want you to build a module that can tell old data from new data. I want you to prioritize the new data over the old data. I want you to take the outcomes and then measure them for degree deviation from some other index, right? And give me the results. And I think that that's where it kind of gets crazy. It's not that I've got your data and I can ask it questions and I can figure it out. It's that I can create on the fly agents that are able to build themselves, build their own interfaces, know what they're looking for, come back and present the information to you all in a prompt kind of logic, right, in a prompt kind of manner. And if you can do that, I think strange things happen. Just like if everyone in the world can log on through a browser and get access to information like social media, whatever, strange things happen to the media world, right? We don't go on to the news at 10 to find out the information anymore. It's on Twitter, it's on Facebook, it's on social platforms that we get the news now. So the world is different, how the world works is different. I think how the world works is different when you have self-building interactions that are coding themselves, that are able to run themselves, that come back and tell you things and then they're able to modify. So you say, if the results don't meet, I want you to rerun the process and I want you to change until this threshold is met and then I want you to review against best practice so I'm building a loop so that it keeps on getting better at the thing it does. And I think that that's the bit that people haven't got their brain around really yet is what changes when that happens. And you're starting to see some of that now with Stack Overflow and the data that was used to train these coding assistants. The use of those Stack Overflows is down something like 50% because people are using the agent now, their copilot, to do it. So their engagement rates are way down and it may be the case that that data now becomes stale or that data only gets useful where it's really edge cases like this particular version of that software and integrating with that particular version of that software and that's the only place that answer is. But the business of Stack Overflow kind of collapses.
VUX World
"paul sweeney" Discussed on VUX World
"And it will result in poor outcomes for you eventually. So I think the answer is it's not that the LLM is great at X or great at Y. I think it's the whole experience. Like saying something like, you'll know what this is, but anyone from probably any other part of the world wouldn't. I'll pay you 20 spawns in a fortnight. And you go, like, what does that mean? It's like, put spawns. What's a fortnight? It's like, I have no idea what that word means. I have no idea what's going on here. So I think that understanding what's being said is one job, like what's being said. What are they trying to do? What's the intent? And then figuring out how to best use that context. Like it might be relevant, it might not be relevant. You might want to use it. It might be noise, might not be noise. Like that's where the next level of experience comes in. And I know that you're keen on this idea of the other data in a conversation that gets shared that might not be pertinent to this, but could be pertinent to something later. And I think that is interesting, but it's a very particular case of do you actually want to know that? Because you're holding that data now on somebody and now you've got like responsibilities around that data. You may not, there's no reason for you to be holding that data. So if I was to say like what's particularly worked, it's the focus on the process and whatever you can use to make that process work better and not designing for, I call it the how fast does a cheater run. Nobody asks a payment service how fast does a cheater run, but they may have many different ways of putting in their address and cracking that problem is probably going to get you that extra percentage completion rate that leads to another thing that leads to another thing. So that's my current view on it anyway. Yeah, definitely. The whole thing about the data that you can get from conversations that could be useful later is an interesting one because there's all kinds of GDPR considerations around that. The example that someone gave to me before was that, let's say you're in the middle of scheduling an appointment and you're trying to book an appointment and you say we've got a Saturday morning at 10 a.m. and someone responds to that and says, sorry, I can't do Saturday mornings. I've got to take the kids to football practice. And the example that was given was, okay, now we know that this person has kids and now we know that they do football on a Saturday morning so now we can then tailor things according to that in future. It's different if me and you are having a conversation and at the end of that conversation I say, enjoy football with the kids on Saturday. You're not going to take that as a violation of your privacy because that's within the context of the conversation. And then maybe even in the future if we manage to speak again but we don't know each other but we're into customer service and we're like, is that Paul Sweeney? Did we speak before? Oh, how are the kids getting them to football practice? Then that's fairly acceptable. But if all of a sudden you receive some sort of targeted messaging about the fact that you've got kids and there's football weaved into it, it's a little bit like when Tariq found out that the girl was pregnant before the parent found out. It's all through data recognition, pattern matching, prediction and then proactive targeting. It starts to get really creepy. Especially when it's done at a machine repeatable level. I think that's going to be one of the questions around what gets remembered and why it gets remembered and how it gets stored and how it gets used. That's the whole point of having data privacy to the fore.
VUX World
"paul sweeney" Discussed on VUX World
"For instance, just outside of us, just buying a software that has a workflow in it, that does things like reminds you to do things and that and that and that, even if it wasn't just... Just putting order on the chaos drives out value in the first place. The second thing is joining up flows from one side of the company to the other. Not that easy to do. So just joining up the flow of data also adds value. Then you get into each of the steps. It's like contacting, like engaging with the people. Then it's like, did I understand what they wanted quicker or slower, or did it take two turns or one turn? Did things drop? The usual conversational analytic stuff. That's its layer as well. But ultimately, it's that experience. If it's a turn or two, what did it feel like? Did it feel like it was easy? Did it feel like it was friendly? Did you feel confidence that it was going to get completed to your satisfaction, and thus you stayed with the process? So there's a number of different things that lead to it being successful. The things that kill you, I think, are where the application says something stupid or asks you to repeat things or doesn't quite catch it, and then you lose confidence in it. So I think that confidence is important and the motivation to be in the process is important. Like if you're kind of looking at something and you're kind of interested in something, that's one type of journey. But if you need to get a bill paid by Friday or you're going to get fines, then you're highly incentivized to stick with something that might not be perfect, and you still get to your end of your journey and you still get good completion rates. So I don't think it's just... I don't think any customer sits back and goes, oh, I really like the way I put five things in there. It was, that's how I do it. I mean, that's my flow. I mean, I've seen people in a message put in two paragraphs into a message and send it in. That's not something that you would expect, but you might get a lot of information in there that you've got to parse through to find out the thing that you want, like I want to pay a bill. But in that message, there's also information like I can't, I couldn't pay it yesterday because I was in a hospital. Like I had an incident with my child and da, da, da. And that, you might go, oh, the main idea is to pay a bill here. The second one is a highly rated vulnerability incident. And then you've got to have a rule to say, well, the vulnerability incident must take precedence. Like you must explore that first before you go into any sort of payment process. And so that's where the detail of a customer experience, like per area, you really have to dive into it. Like you might just go, oh, take the money and move on. That'd be the wrong thing to do.
VUX World
"paul sweeney" Discussed on VUX World
"But we're still in the early days. These are still early days of using agent assist. It'd be agent assist now. If you think about the customer, the customer on their web chat, they're getting customer assist. It's like the customer's asking questions, can that web chat window help them get their job done? And so I kind of think that before the LLM hit, so to speak, and we're in that trough of disillusionment about conversational AI, I said, look, the thing that's on every single website is that little button in the corner. It's that little chat button. It's nearly on every web property in the world. And customers know to click on it to get something done. So the digital assistants are still, they're going to be called many things like digital assistants, custom digital assistants, intelligent assistants. It's going to have a range of different names to them, but essentially they're on your web property. So how can I make my web property work better for me and get the job done for the customer? And that's where it's going to be. That's where the creativity is going to come in, like buying a car. It's going to be different than buying insurance or paying your bill or doing something else. Yeah, definitely. There's a lot of good conversations going on here in the chat. Thank you all for chiming in. I won't read them all out because I think some of it was related to some of the previous conversations we were having. So we covered a few things there. One of the things that you mentioned is definitely I agree with, which is that not to look at where the technology is today, but think about where it can get to. And I think that for the first time, whereas you use the analogy of Alexa or something you talk to, the thing with Alexa and all of these voice assistants is that they're all predicated on the intent based model. So they're all trying to put your question into a bucket. They'll have a hard time at doing things like compound intents, which is where you'd say two things at once. I need to set up a payment arrangement and I also need to pay something this Friday on another day. Two different things that you need to do. We also have a hard time with ambiguity. As you said there, if someone says my holiday is just finished, are they talking about the actual holiday and therefore now they can pay because they've returned from holiday? Are they talking about the holiday that they were on? The now payment holiday is not finished. They're not great at disambiguating like that. So I definitely think that large language models have a role to play there. Also, when you move out of intents, for example, if you were to put a conversational UI on top of software, as you alluded to, that was the thing that we were thinking when we saw Alexa. First of all, this can be used to navigate anything. Any piece of software, any system that has any buttons or you need to take any action could fundamentally be boiled down to an utterance that you could say and then get things done quicker. So I can definitely see all of that occurring. There's definitely places where it can serve a purpose now. There's lots of companies who are using large language models for document summarization and even response generation in some parts. Not that many of them are doing it publicly facing, but some are. A lot of them are using it internally. So there's definitely value there. The thing that you went on to say, which was the increase in payments that you received, increase in income expenditure forms being completed, 50% increase in payment arrangements being set up, 70-something percent increase in actual payments being made once they're promised to pay. Do you put all of that down to the experience that Webio has built up over the years to refine those conversations, to build those language models to understand people effectively, or do you put that down to incorporating large language models into that equation? Yeah, it's hard to... The large language model, is it because of the fact that you've worked on this problem for so long that you've got it pretty much nailed? I think it's a combination of the bits.
VUX World
"paul sweeney" Discussed on VUX World
"So I think the point about throwing, we might just shift a little bit into the future stuff, is I'm pretty sure that every company is going to have an LLM. I think every company, every enterprise is going to have a large language model or a language model or a language model type technology in the middle of its company. And I think it'll look something like infrastructure. So what will it do? Well, you're going to have all your documents indexed, you're going to have all your interactions indexed, you're going to have every conversation that everyone's having in the company indexed. And then it's like, okay, now what happens? Is it prompting? What's the interaction style? How do people get value out of it? I do try to give credit where credit's due, and I've used this example before, but it's nicked directly from Brett, which is people, when they had their Alexa, it was in front of them and they spoke to the device. Now the assistant's sitting next to you and you're working with it. And so I think we've been saying, everyone will have a co-pilot. And I think that's a great brand for it. Everyone is going to have a co-pilot. Every job will have a co-pilot or multiple co-pilots. And if you ask engineers, developers, what are you using as your assistant now, they all use your companies. They all have their tools to help them. If you look at any marketing team right now, they are using this technology to develop letters, blog posts. Every piece of content they can develop with this, they are already using it. Whether it's signed off to use or not, they're all using it. And so I think what you're going to see now is the finance people will get their hands on a couple of neat, hey, interrogate this Excel sheet for me, change the formulas, make it do this, make it do that. And then they'll go, holy crap, this is really kind of powerful stuff. And it'll just accelerate and accelerate and accelerate until it's kind of normal for you to figure out, oh, what are you using for X or what are you using for Y? And they'll all be tied back to some form of large language model or language model technology in the middle. So I think that I'm not so today caught up in will it be approach A or will it be approach B that kind of wins out in terms of a deterministic conversation or flow versus a completely LLM driven flow. I think ultimately it's just can it be validated? Is it privacy first? Does it meet the standards that the industry are operating in? And can you point to the value that it's creating, the investment and the investment in it and what it's delivering? And so one of the benefits of being a focused player is we're able to look at things like, oh, just to take, again, a particular niche example, there's a thing in our business called income and expenditure reviews, figuring out what the income of the person is, what their expenditure is, how much disposable income is. And if you do that on the phone, it can be anywhere from 20 to 40 minutes, depending on the age and time. And so if you can do that with a conversational interaction with an automation and that can be done, obviously the time from the customer might be a little faster, but there's no age and time involved in it. So it's much more cost effective. And what we got is a 60% response rate from customers on completing those income and expenditure forms. And if you know that business, you go, well, that's phenomenal. That's a really good completion rate. So the engagement rates for these things are very good. And then the outcome rates are things like you're trying to get someone to do a promise to pay, say, I'll pay you this Friday. If you get a promise to pay, then they do tend to pay. So we've got an increase in promises to pay and then a 52% increase in payment arrangements and then a 76% increase in total payments. And so across a business, these kind of numbers are just massive.
VUX World
"paul sweeney" Discussed on VUX World
"I think that that's, like, just for the part of this conversation, which is about this is all very well. It doesn't work exactly as I want it to work today. I go, yeah, I just put a little step in here and go back in time. If you go back to the year 2000, 2002, or even three, I remember being, because I think we're in a shift. We're in a shift that is like as big as mobile and as big as cloud. And I remember someone explained to me how computer software worked and was, you know, okay, you buy these disks, and they come, they get sent to you, and then you have to install them, and then you've got to train people. And then, you know, 30% of projects actually worked at the end of the day. And then there's all these consultancy fees, and I said, God, this is crazy. If I could just open a browser and use software, and it didn't work, I'd just go to another URL and find that software, and then that worked. That's what the future is going to look like, and there won't be all that other stuff. And I remember standing in front of people in 2003, and they're telling me no business is going to put their software in the cloud, and nobody's going to make a payment through the cloud. It's just never going to happen. Because they just couldn't see how the disruption was going to come through. Now, the difference here is the disruptions happen very quickly. But back in 2000, the idea, I worked at Telco at the time, and the idea was if you are in a Telco, the thing you were thinking about was faster broadband is going to bring faster video, and your competitor is going to be the cable company. And the cable company is going to give you your phone and your broadband and your TV. And as a Telco, you're only going to be able to give them connection, so it's like those huge amounts of fear. And so you were thinking, broadband is going to bring faster video, right? It's going to bring TV, like streaming TV into everyone's home. And actually, in 20 years, probably the biggest impact of broadband is we're able to work remote. And then working remote means that as someone, I've worked remote for 20 years now, and I've been here at home with my kids, been able to collect them from school, they have a relationship with me as a father that they would never have had before, that maybe fathers were never able to have before with their kids. And that's a huge social change. And now, post-pandemic, we're seeing, oh, hold a second, we don't have to hire engineers that come into Dublin City Centre and have to get apartments in Dublin and have to live in Dublin. We can get engineers from anywhere in the world and work remotely with some fine talent. So it changes the nature of society, it changes how we work with one another. And I think that's what's going to happen with the AI. So while it's not everything today, it's doing some pretty neat things. But the important thing is, what's the trajectory and what is exponential and what are the new things going to come out of the middle of it? And they're not always obvious.
VUX World
"paul sweeney" Discussed on VUX World
"And also, you've pointed out that Webio is an award-winning solution, actually. Thanks for reminding me, Brett. I should be better at blowing our own trumpet. But because of the focus, we have genuinely won every single customer experience, best use of technology, best use of AI, best use of ethical AI. We have a cupboard of awards. And for a small company, we've beaten some billion-dollar companies in these competitions. So we call it the power of focus, just focus, focus, focus. So you touched on something interesting there, though, which was that you began... So Webio now is essentially a platform, but also it's a very domain-specific intelligence. I would compare it to like a Hyrule or a Euphonia, which is like... So in those cases, it's like they would handle a customer journey around operations, booking of an operation at a hospital, follow-up care, making sure everything's all right. So the whole customer journey around that operation is being taken care of by a Hyrule or a Euphonia. So if a client was going to bring one of those companies on board, they would just essentially procure the solution. And it's all pretty much already done, because the conversations are all the same. The only thing that's going to be different is the data that you feed into it, because you're going to have different systems. So it sounds as though Webio is kind of like that, where a credit or collections company close to Webio, you provide these use cases. It might not be out of the box from the perspective of just turn it on and it works. Maybe you might have to do something to integrate into a different system or some kind of thing like that. But most of the heavy lifting around the conversational kind of like design, the technology, the stack around managing those conversations is kind of set in place. Yeah, it would be very similar. But again, we find that customers always want to change something. There's always something, but nothing radical. And this is different to theirs, kind of thing. Yeah, and some things like, let's just say someone says, if the card is declined, I want you to offer another card. There's another card on file. It's like, oh, okay, that's a nice thing. One card could get declined for any number of reasons. There's a second card, just put that in the process. Now, if you try to do that with a regular kind of off the shelf solution somewhere, they go, oh, no, you can't add another card. That's just not something you can do. So adding gateways, how to treat the gateways, that's another element of the flow. Again, we'd be familiar with that. Yeah. So the reason for wanting to clarify that is that a lot of the logic around the conversations are kind of built, so to speak. So if it's somebody who needs to make a payment, then there's a conversation to be had around making the payment. If it's somebody that can't afford to pay, there's a conversation to be had around not being able to afford to pay. And although there may be differences on different clients, the pattern is kind of similar. So you must have this kind of logic. You mentioned there the NLU works in tandem with the workflow and stuff like that. So this workflow is deterministic, basically. But then you add large language models into that. And so now I'm kind of wondering, because this is exactly the situation that a lot of companies are in now. They've got a deterministic NLU and workflow system, built on Watson, built on Dialogflow, built wherever. And now there's large language models. And on the one hand, you've got a group of people that are like, throw all of that stuff out because large language models have got this covered. On the other hand, you've got people who are like, well, actually, no, no, it needs to be deterministic and large language models are going to supplement. So just curious about how you've managed to integrate large language models into that process that's already working. Where exactly are they featuring? So it's interesting to see what's value, what's exploration, and what's future. So right now, picking up things like that's somebody's name, that's someone's address, that's a number, that's an order number, and this is what we want to do with it. That's as far as people want to go. And if you say, oh, I could run that through a large language model, but it's going to be wrong a couple of times. But millions of iterations, who cares? Well, companies care. Companies do not want to turn everything over to a large language model. It's just not going to happen today. It's just not going to happen today. Now, that's not to say that it won't happen in the future. I'm actually fairly bullish on large language models being able to handle a lot more in the future. But today, you really can't do that. And I think you especially can't do that in a financial environment or a financial and credit environment. A financial and credit environment where there's potentially vulnerable people at the end of that process. You have just got to be laser focused on being compliant the whole way through and being demonstratively compliant the whole way through. So that's the first thing. If you want to try the leading edge of using mid journey and open AI to develop some really fascinating marketing materials to knock your lights out. But you will not have a company turn open AI to say, generate me a response to this message and push it out to a customer without a human agent having the rise over it at the very least. So the environment that you're in will kind of dictate how much you can do there. If you're in marketing, you're having some fun leisure company, clothing company, you have more latitude to do some interesting things there. With regards what people are getting value out of, this might seem like incredibly stupid, right? But here we are. If I put a message into a system and I go, hey, this is my name, I'm contacting you about this bill, I think it might be too much. It says 150, but I can pay you 50 today and 50 next Friday. Is that okay with you guys? Right? That's like a bunch of information. If you were able to respond to that person by going, hey, Paul, that's fine. 50 euro today, 50 euro next Friday, all good here. Just press yes to execute this deal, whatever. I go, yeah, cool. And off I go. Now, you go, that doesn't seem like a long, interesting dialogue with the avatar realistic real time. But that's not the job, right? That's not the job. The job is customer has a question about how they're going to pay a bill, change a bill, change a plan, move to the next thing. So I think that focus isn't going to change for customer service for a while. Now, if you want to just shift a little bit there and go, well, what if I wanted to design a new program? Let's say we're interacting and through the course of this interaction, I'll let you know I've got pets. And you go, oh, your pets aren't on your insurance. And you go, oh, can you put them on my insurance? And you say, I'll generate you a policy that has pet insurance in it and I can sign off. I can see that kind of a thing being dynamically produced products or dynamically produced services as something that will be useful in the future, especially if they feel really natural and they're kind of compelling and they've got some legal status at the end of it. So I think that'll be interesting. I just think the idea of the completely deterministic flows versus probability flows versus putting in something else, like a complete LLM driven flows. I think that the LLMs will get good enough to do it. I think the question is how auditable they are and how much privacy they guarantee in that process at the same time. But I don't see that being that far away. Like, I'd be surprised if in a year's time you weren't able to do kind of prompting of chain of thought and processing steps and have those validated at each point and have processes for validating that off of the LLM so that you can do some validations. But in the meantime, I think I'd just be a little wary of it. So am I understanding it correctly then? So your recommendation is not to use it for the production of responses that go out, so not to generate responses. However, maybe to experiment with it in that use case you said where someone says, I've just come back from this, I'm doing that, I can't afford this, I can do 150. Is that what you're saying? Again, this is where the devil's in the detail, right? So you could have an LLM generate a response and have it very accurate. What I'm saying is I would not push that out to an end customer without a supervisor or human agent clicking on it and going, I've read it, that response is fine to me, go. So that's one thing a large language model could do is to generate the response. The other thing it could do is to use it to understand what's been said. For sure. So not necessarily responsible for generating, but maybe responsible for classifying. So is that the example you were getting at with that long utterance with lots of entities around pricing and dates and so on? Or are you still using a more traditional NLU? I think you can do both. Again, it's mix and match. I think you can mix and match there. I think just before you get onto the kind of tagging and picking up long tail intents and stuff like that, just goes a bit before that on summary. So like summary conversations are very interesting because summary, and this isn't, again, this isn't completely new to the world of customer interaction and LLMs right now. So if you look at a conversation and you go, what's the value in summarizing this conversation? Well, there's a couple of bits in that that might be very valuable in a summary. Like you could say one agent has been in a back and forth with a customer. The agent needs to pass the conversation on to another agent. The agent gets the conversation passed on to them. The summary comes up and you go, oh, this customer contacted us about our returns. They're interested in also getting a track suit. They're just looking for a pricing and can they get a discount? That's the summary of the conversation. And it's summarized in the context that I'm sure there's lots of other things that you could summarize in it. How are you summarizing it in context for that particular, like what priorities do you have? Like you might, you could have also said this customer is not yet happy with their returns because they haven't finished the returns process. They haven't expressed that they're happy with it. And you might have just prioritized that return, that return part of the conversation. So summarization has a bit to go yet. Right now, it's just saying like, we love this idea of dropping the documents in and going, summarize this document for me. Right. And it goes, hey, this document is about X or this book because the context window is so big. Now we can just drop a book in and say, give me a summary. So what I would say is that I think there's more to go in the summarization piece. And I think it's, I think there's these kind of trade offs still in the technology. So I've been following it. So I think what I would say at the moment is the long, the big, the bigger the document, the bigger the file, the more likely it is to have hallucinations at the moment. If you can break it into parts, the more accurate you get at each part, and then you've got to recombine the parts to get another level of summary. So I think there's different techniques that would be applied to it. But fundamentally, what are you trying to do? Like, what's the job to be done? And it could be in a summarization context, it might be, tell me what the conversation is about. And then give me an assessment of where it's likely to go. Like, how is it likely to end or what's going to make the customer happy level A, happy level B. Are you able to move the conversation in one of those directions? So I think, and just final point about summarization is, summarization is a summarization, but it's also context. So if you're able to create a really good summarization, you're also creating a really good context. And when you have that context, now you can point it at other resources and other things to say, oh, actually, given where we are in this conversation, this person probably needs this article or this returns detail or some other, like, because you've got the context, you can more accurately find the piece of information or action that they would want to know. So summarization could also help with that context, I think. Yeah, we've been using summarization on a lot of things from podcast transcripts to video transcripts, getting social media posts out of it and stuff like that. And it's, you know, it's hard, it's better when you've got a longer bit of text, kind of, but some of the research I've seen is that if you've got a really long piece of text, then it works better at extracting information from the beginning or end of the text and it's fuzzy in the middle. You kind of use it for podcast transcripts and we'll just say, like, pick out three themes that have happened in this thing. We've got three themes and then we'll say, okay, we'll write a 500-word article about that theme. And that kind of works okay, but the summarizing of video transcripts is less effective. It just basically, it'll take like, I don't know, three, let's say, 500 to 1,000 word video transcript, and it will almost just pick out lines from the transcript, reword the lines and then put them back together into like a paragraph. A lot of the time, it's not actually that good.
VUX World
"paul sweeney" Discussed on VUX World
"And I use a couple of examples to say that the nuance to this might be a little hard to pick up if we're not in the industry. But if somebody said to you, I can't pay that bill, I'm just off my holiday. And you go, you're telling me you've gone on your holidays to Corfu and you're coming back anytime you can't pay your bill. You know, how dare you? It's like, no, actually, in this context, the holiday was a payment holiday and you were given a payment holiday because of circumstances you were in. So treating that word holiday in one way means one thing in one context and one thing in another. If you said something like I can't pay that bill at the front of a conversation versus at the back end of a conversation means two different things. So at the front end of the conversation, I can't pay that bill might be I do not have the financing to pay that bill. I don't have the available funds. At a later stage in that conversation, it would be, oh, I can't pay that bill. The gateway isn't working. It's not taking my details. Something's happening here. So just where it is in the conversation can change the meaning of what's happening. And that's on top of all the regular kind of lexicons and steps and approaches that credit and collections companies take. So where we are today, like from where we were before, like if you think about just what I called earlier low context, low context SMS, like yes, no, will, can't. It's like small words, small sentences. You didn't need a whole lot of sophistication to do that. But try doing that at scale. It's really hard to do that at scale and really hard to do it over SMS at scale and then leveraging it into the WhatsApp and then leveraging it into the web chat. And so now what you have is the ability to run, you know, straight like credit and collections conversations over messaging, over a web portal, like over forms, over other containers. That's how I kind of think of them. It's still the one conversation is just happening in different containers. And at the end of the day, what our customers want is they, it's kind of this double thing. Like on the one hand, you want to be able to collect your money effectively. You want people to pay their bills on time or to make arrangements to pay their bills. But you also want them to be able to do it in a kind of friendly way. You want to do it in a way that's supportive, that's good experience. And you want to do it in a way because of the particular area we're in, you want to do it in a way that has one eye open for people that may not be able to pay a bill or shouldn't pay a bill because it puts them in difficult circumstances. And that's the whole area of vulnerability and duty of customer care, duty of care, which is now coming in across the UK. And that means that you really need to show the regulator that you have processes in place where you're actively managing to find out and sense are people in difficult circumstances? How much you might prove that and then how might you demonstrate to the regulator that you are doing this in a professional manner? And so that's where we're finding a lot of interest as well. Does that give you a flavor of it, Cain? Yeah, absolutely. Yeah, yeah, yeah. We've got some comments here, Mr. Breckinsella, shout out to Brett. It's nearly impossible, going back to what we were saying before, it's nearly impossible for service companies to pivot into products, the operating models, philosophies and customer experience in conflict, agreed.
VUX World
"paul sweeney" Discussed on VUX World
"So that was the deeper vision like that, that people would need their own kind of control of their own AI. Now, that's evolved again, before going to the evolution of it, just a couple of key little bits that the viewers might want to know or need to know. So that goes like that idea of like, we would want to be able to give you your SMS, we want to give you connection to your WhatsApp, we want to give you connection to your web chat. And we want to control that journey as much as possible for your, again, session management security and the evolution of the overall product. And then you would get some agent, again, agent screens, managing queues, queue management, conversation management, bringing data into the conversation so that agents can use it. Nothing that you would, nothing too dramatic there in terms of what you would expect. And then like the next kind of where we are now is right. So with your own AI, trained on your own intents and trained your own data, you can bring in AI into the conversation and say, I want to pick up things like, oh, let me see, like, what would you say? And not just a custom entity, right, a custom intent, custom values. You can then pick those up and maybe look back, use our API, look back into your own solution and go, hey, this means that I need to know what the current balance is, or what the current future data payment is for this thing or some other value in your database and bring that back into the conversation. And then we've built our own natural language understanding engine and workflow that says, okay, we're now able to say that value is above or below some value or percentage, say yes and move to the next step. And so it's an example of, because we give you this really, like, I know companies say they've got the best of everything all the time, right? But the API that we've built for, like it's built for credit collections, getting your data into the conversation is your leverage. Like, you will tap out unless you can actually get into the enterprise data. So we spent like a really phenomenal amount of time and effort making this a really robust, really usable API. And then you bring that conversation into the data, sorry, that data into the conversation, and we've built our own Webio flow. So again, it's workflow and NLU. And we have this kind of approach to building what's called a custom language model. So it's using the LLM technology. It's all based on open source projects that we bundled together and made our own. And then we've taken conversational data from real conversations and built a custom credit collections language model.
VUX World
"paul sweeney" Discussed on VUX World
"And it dawned on me on that platform that every service was going to look like that. Every service was going to feel like that. And in 2016, there was an article published by Chris Messina on Medium, which was 2016 will be the year of conversational commerce. And he showed this example of two people speaking back and forth or texting back and forth on WhatsApp. And they say, hey, would you want to meet up in town? Yes, I do. And the app figured out that, hold a second, Paul is on this side of town, the person he's speaking to is on that side of town. If they want to meet, they're going to have to get a cab. So it would insert a link into your Uber and say, hey, do you want to get an Uber to that location? That's it. It's like a halo, but in your messaging. And not only could you get a cab, but you could pay for something, you could find a pizza, you could, whatever you wanted to do, you could do it all in this messaging. So it just hit me like a bolt of lightning that this is exactly the way we're going to interact. Now, it all didn't work out like that at the end of the day. But I was convinced that we would be able to call up services in a messaging service, we'd be able to interact, share things, negotiate things, and we just do this in this kind of messaging environment, which was like a super app. That's what they call it now. And so once you see that shift, we wanted to be a part of that. We wanted to get into that conversational commerce kind of space. But it was obviously very early, it was super early. So we said, what we'll do is we'll start out with SMS conversations, because we know that people spend money, companies spend money on SMS conversations today. Even if the complexity handling isn't there in the background for these new apps at this point in time, you know, low context conversations could really be handled pretty easily on SMS. And as I said, we had some experience in how SMS worked. So we targeted SMS conversations to start with, and the idea is that we would migrate then people and companies into WhatsApp and into other messaging apps. And at that time, it looked like there could be like 10, 20 different messaging apps for people. But that's really become WhatsApp these days. I mean, WhatsApp and some Facebook Messenger. And then it just kind of spiraled from there into what you could do. And then, of course, the hype cycle kind of started to kind of come to its peak. And we went, you know what, we really need to focus here. Like we can't be all things to all people. We need to focus. And so we saw some pretty good results with conversational SMS in the credit collection space. So we just laser focused on that and said, you know, from here on in, we're just going to focus on credit collections conversations, and we're going to be the absolute best at those. So we're not going to be an average player across, we're going to be the best player in a niche. And from there on, and then the direction of the company is kind of being directed by the needs of that particular market. And that continues till today. Am I right in thinking that way back, that Webio was kind of like a consultancy that then went into product? Or were you always focusing on product first? Always product first. Yeah, there's a number of different, I mean, there are, maybe this has changed in the last couple of years. But it's actually fairly hard for service based companies to pivot into product, because it's a matter of priorities. And it's a matter of, you know, somebody says, like, Oh, look, I'll give you 30 grand in consulting to get this integration work done and get this working. And then your revenues on your products are two grand, like you're going to focus on what's bringing in you 30 grand, that's human nature. Even if you're in the long term, the value might be in a product. But you can't afford to do that in the short term. So you have to focus on the services business. So it's actually very difficult job to go from services to product. Yeah, I can imagine that definitely. So you began working, trying to figure out what a product looks like, you're focusing on SMS, because you kind of a really kind of big believer in the conversational commerce component of conversational AI, you start working with clients, you get some traction in the credit and collections kind of space. And then you begin building a product around that. Yeah, tell us where weather your is today, then because that was, I think you first started focusing on that in like, what, 2018 2019, something like that. Yeah, yeah. So it's, it's a good space, right? So it's good space of time. I mean, so. So where we are today is we to kind of interacting with space kind of went. Look, I think what what we need to do here is concentrate on being like, absolutely the secure service, your data doesn't go anywhere. You're not using outside services to make this service happen. You know where your full data journey is. And so what we want to do is give you the full stack where we want to give you like soup to nuts, everything you need. And because we control that, we're able to nuance everything in that stack over and over again to give you the service that you need. So, as you know, when you get like from the middle, like lower, middle and upper market, when you go to the bigger customers, they always need things like they always need like, I don't need it to work like that. I don't need I need this change to need that label different that shouldn't be like that. And if you, you'd be surprised once you go digging in, like, you can't change that, like that comes from Google dialog flow, or you can't change that that comes from somewhere else. And even simple things like you think there are simple things, like moving from one service to another, and passing through values and passing through details. Not at all simple when you really get down to it. So we were, we knew that we wanted to be the secure company, the one that you are perfectly comfortable with, where your data is going. And also, we were very early into the ethical AI sphere to say, okay, we want you to be able to say, your AI is trained on your data. If you want to label something one way, you want to treat labels one way, that's entirely like that's you. And if you want anything in the rest of the application to interact with the AI in a specific way, then we can make that happen for you as well. But only if your AI, if you have your control over your AI, and that was like core to everything we've done all the way up, like with how the data is imported, how it's anonymized, how it's trained, how it's auditable, all that stuff.
VUX World
"paul sweeney" Discussed on VUX World
"I think it's fair to say that a couple of ideas have been percolating in the back of my brain around communications and particularly asynchronous communications. So I've been involved in a couple of two-way SMS experiments and programs, and what we realized is that there was a lot that could be done with SMS that wasn't being done, that could be done. And a lot of work that you would say people thought that could only be done by voice, like outbound calling and agent-based communications, could actually be done fairly well with SMS. With the real kicker in it was when you go with SMS conversations, you save an awful lot of agent time, and agents were able to deal with a whole lot more conversations. And so that was a huge productivity gain. And then we noticed that the... I was following figures for SMS traffic in Europe, and at the time, one of the Belgian telcos was putting up numbers saying that their SMS traffic was really decreasing quite rapidly, and they put it down to a new service called WhatsApp. And so WhatsApp hit the radar and was looking at that, and then the third kind of thing that happened was there was an app called Halo, and we were calling up cabs and getting cabs around the place. And what I found... I was going up and down to Dublin a lot, and what I found was that all the anxiety about me getting a cab and getting to my train and getting home to my kids in the evening on a Friday, there were two small kids, I wanted to read them the stories in bed, missing that cab ride was just a big pressure for me, especially in Dublin. And when the Halo came along, I was able to get the cab, see the cab, see it on its way, know the ETA, get out of the cab, give them the pay instantly in the app. And it was just such a relief. And one day I did not get that cab. I just stumbled on the side of the road and I flagged the cab. I got in the cab and it was like the cab was filthy, and the driver was racist, he drove me the wrong way to the station, I was fumbling with all the change to get out of the cab to catch the train and get a receipt. And I was just standing on the platform at the end of it and I says, Oh my God, I'm never flagging down the cab again. I'm never doing that again. And I realised that it wasn't just an app. It was a whole experience that people behave differently in the cab. You are polite to the driver, they're polite to you because you can rate one another. Like you've more or less of a chance of being picked up next time based on your rating that the driver gives you. And they need to be nice to you because your rating of them means that you'll use them again or recommend them or their rating goes up as well. So the experience was completely different, even though the whole thing was this kind of seamless, frictionless technology, the actual effect of it was an experience and I was just obsessed by it.
VUX World
"paul sweeney" Discussed on VUX World
"Yo, yo, hello, ladies and gentlemen, boys and girls, welcome to Vue UX World. Fresh off the heels of the Unpassed conference that happened last Monday and Tuesday. Not the one just gone, but the one before that. And it was, according to some of the people that were there, the best conference that they'd ever been to in their life. And so, don't take it from me, that is just passive. I'm just channeling the feedback of some of the people that were there. It was absolutely immense. We had two full days of epic programming. We had conversation designers and some of the who's who of conversation design done upon London, and it was absolutely amazing. We sold out, had an absolutely full house. There were some very nice networking opportunities in the not entirely sober afterparties, but it was absolutely fantastic. And so, we will be back with Unpassed 24 on the 17th and 18th of June next year. So, yeah, thank you, everyone who attended. And those of you who watched online and for those of you that didn't manage to get a ticket, then, you know, keep your eye out over the next few months when the 2024 tickets will be released. Because it was immense and next year is going to be even better, if you can even believe it, which I can. And before I introduce today's guests, I want to give a shout out to Tideo. Tideo is an AI powered customer experience software platform. It has live chat and it also has built in kind of automated conversations. It's specifically focused on the kind of small to medium sized businesses that are predominantly online businesses. Think about retailers and things like that. It already has out of the box pre-designed and pre-configured domain specific use cases. Things like product recommendations, things like product availability, shipping questions, order status, returns, conversations, all that kind of stuff. So, if you are a small to medium sized business and you are wondering what this whole AI revolution is all about, you can get started with Tideo very, very quickly and very, very effectively. It's answering four out of five questions that their customers have. And so you can basically start automating a lot of your kind of initial sales and products and customer support conversations by using Tideo in a very short amount of time. So you can visit tideo.com forward slash V-U-X if you are interested in finding out more. I'll put the website on the screen now. For those of you watching online, for those of you tuning in on a podcast, it's T-I-D-I-O.com forward slash V-U-X. If you do go to that landing page and you do sign up, then you will save 20% on your subscription. Now, you can't say fairer than that unless, of course, you can't pronounce your T's or your H's. But there you go. Tideo.com forward slash V-U-X. Thank you, everybody. And thank you. Thank you, Tideo. All right. Today's guest is, I don't mind saying that he, I would say, is a legend in this space. I feel as I've introduced a few people with the title legend, but Paul is definitely, I don't bandy that word around freely. Paul has been deeply involved in the conversational AI. Paul Sweeney has been deeply involved in the conversational AI community. For a long time now, Webio is absolutely smashing it. Conversational AI platform that began servicing debt collectors. It's now expanding into all kinds of other industries. They've got some very special technology. We might learn a little bit about that today. Paul has got some great perspectives and immense experience in conversational AI and conversational automation. And so if you are interested in how do you get real value from this stuff today? What are the value of utilizing large language models? Where are they good? Where are they not so good? What are the benefits and results that you can achieve when you implement this kind of stuff successfully? And we're going to have a little bit of a conversation about what the future holds. Let's imagine that this technology continues to be adopted wildly. What does that mean for your business? What does that mean for data? What does that mean for individuals and the productivity, power or risks that may be involved in implementing this kind of technology? So, ladies and gentlemen, please, without further ado, please welcome Paul Sweeney. Paul, welcome to Webex World, my friend. Thanks, Cain. I feel like I should applaud myself after that introduction, will you? Thank you. I've had a long time to rehearse that because it's been a long time coming, this. I mean, we've known each other for a long time, and I can't believe this is the first time you've actually made it onto the podcast. So apologies for not doing this sooner, but thank you for your time. Oh, I'm delighted to be here. And I will say publicly, on record, that Paul saved my ass in March when we had a couple of speakers pull out of the event we were doing at the European Chatbot Summit. We did it, and the event turned out to be absolutely spectacularly successful with the help of Paul, who dropped everything at short notice, put together a presentation, got on a flight to Edinburgh and came and presented and was part of the event. It was absolutely fantastic. So thank you again for that. I'm such an opportunist.
Bloomberg Radio New York
"paul sweeney" Discussed on Bloomberg Radio New York
"World of national news with rhythmic. Paul Sweeney, President Biden speaking out again about OPEC and the announced oil production cut. Disappointment and we're looking at what the alternatives you may have. President's not being specific about those alternatives, but he also says he does not regret his recent trip to Saudi Arabia where he asked leaders to keep crude flowing. As the president's special envoy for energy security, he says they're focusing now on production at home. We have had conversations with the leadership of almost every major oil producing in the United States, told them, what do you need to instead be incentivized to increase production? Almost seen at The White House tells Bloomberg television. There is also more work to do on the strategic petroleum reserve. Russian troops killed at least two people in rocket attacks in the city of Zappa region that's about 32 miles from the namesake nuclear plant, a regional governor says high rise apartments and other civilian targets were hit. Julie Norman at university college London says this only adds to the list of Russian atrocities in Ukraine. Obviously it's hard to verify all information all stories that are coming out in a conflict situation, but we have seen just so many patterns from Russia's atrocities in this war in different parts of Ukraine. University college London's Julie Norman spoke on Bloomberg surveillance. The death of a 16 year old Iranian girl is adding new field of protests that have rocked that country for nearly three weeks, nika is shock arami went missing on September 20th after she joined the demonstrations against the death of 22 year old masa amini, her aunt tells BBC Persian, they identified her body this week at a detention center morgue, Oslo based Iran, human rights says at least a 154 people have died in the protest. Global news, 24 hours a day. On air and on Bloomberg quicktake powered by more than 2700 journalists and analysts in more than a 120
Bloomberg Radio New York
"paul sweeney" Discussed on Bloomberg Radio New York
"Podcast with Paul Sweeney and Matt Miller Are there some sectors that you want to have more or less exposure to We've got a vaccinate the whole world analysis of the day's Wall Street actions Bloomberg opinion and influential newsmakers The bond market was the boss Bloomberg markets with whole Sweeney and Matt Miller Subscribe to stay at Bloomberg radio dot com The Bloomberg business app or iTunes Before and after the pandemic Is it driven by politics or by science That may be how we keep track of our lives from here on out What do you think the political effects of that impatience will be And through it all there's been Bloomberg We begin on Capitol Hill the most accurate business world and healthcare news before and after The fundamentals do not justify this price action Bloomberg radio the Bloomberg business app and Bloomberg radio dot com Bloomberg the world is listening Would you say it's more important to gather information fast or to have it first or to be the most accurate They can really move the needle when it comes to programs It's the level of support you get from companies for this What if you don't have to choose It also has enormous importance for the labor market How do they get ahead of different administrations We see this move towards digital currencies Bloomberg radio the Bloomberg business app and Bloomberg radio dot com Bloomberg the world is listening We used to take our freedom of movement for granted Not anymore It's not just that people work for the airline and it's natural to feel grateful for the things that kept you going Does America have a chance to lose our advantage Can we get to her community fast enough so that will be in good shape But really we were just doing our jobs Oxford University is starting a study on patients who've recovered from COVID Bloomberg radio the Bloomberg business app and Bloomberg radio dot com Bloomberg the world is listening United there.