WEBVTT - Global IT Outages Creates Chaos and Netflix's Mojo Is Back 0:00:02.440 --> 0:00:10.280 Bloomberg Audio Studios, podcasts, radio news. 0:00:08.720 --> 0:00:14.000 From Marhart where Innovation, money and power Collie in Silicon Valley, NBN. 0:00:14.360 --> 0:00:17.880 This is Bloomberg Technology with Caroline Hyde and Ed. 0:00:17.920 --> 0:00:18.640 Loved Love. 0:00:31.880 --> 0:00:34.080 Live from New York and San Francisco. This is Bloomberg 0:00:34.080 --> 0:00:38.680 Technology coming up full coverage of the massive IT failure 0:00:38.920 --> 0:00:41.839 that is grounded flights and upended businesses and markets. 0:00:42.680 --> 0:00:46.160 And Netflix gets its mojo back strong, so subscriber growth 0:00:46.280 --> 0:00:48.519 extends its lead on streaming rivals. 0:00:48.760 --> 0:00:51.919 And in an exclusive interview, Mark Zuckerberg calls Trump a 0:00:51.960 --> 0:00:56.080 badass without endorsing him for president. More on politics and 0:00:56.120 --> 0:01:00.160 take ahead, but first the market ramifications of the eight 0:01:00.520 --> 0:01:03.080 issues that have hit across from Asia to Europe to 0:01:03.120 --> 0:01:05.759 the US. And it was during European and Asian training 0:01:05.760 --> 0:01:07.959 that we hit the hardest. I'm looking the NASDAK futures 0:01:07.959 --> 0:01:10.039 therefore ed because actually we're only trading off by about 0:01:10.040 --> 0:01:12.360 three tens percent on the major benchmarks, but we did 0:01:12.440 --> 0:01:14.800 hit our loads at about four five am. That was 0:01:14.840 --> 0:01:18.679 when we really started to worry about the global ramifications 0:01:19.120 --> 0:01:21.800 of one key company ed that you're watching. 0:01:22.280 --> 0:01:24.959 Yeah, and it's CrowdStrike, and CrowdStrike is down big a 0:01:25.040 --> 0:01:28.840 market leader in endpoint ransomware security, put out a faulty 0:01:28.920 --> 0:01:33.160 patch or update and that impacted Windows hosts around the world, 0:01:33.280 --> 0:01:36.440 didn't impact mac os, didn't impact Linux if you're in 0:01:36.480 --> 0:01:41.440 that camp. But separately, coincidentally, there was also a Microsoft 0:01:41.520 --> 0:01:45.120 asu're outage, both of those contributing to the downside in 0:01:45.200 --> 0:01:46.080 Microsoft stock. 0:01:46.600 --> 0:01:48.160 We have a number of updates. 0:01:47.840 --> 0:01:50.120 There have been fixes issued, but so many of you 0:01:50.160 --> 0:01:52.640 have reached out to Karen I on social media to say, yeah, 0:01:52.680 --> 0:01:54.920 turning the computer on and off fifteen times seems to 0:01:54.960 --> 0:01:55.400 have worked. 0:01:56.880 --> 0:02:01.200 Blue Screen of death as it was no throughout any 0:02:01.440 --> 0:02:03.880 airport that you're unfortunately stuck at. 0:02:04.040 --> 0:02:05.960 But we want to take a closer look. 0:02:05.880 --> 0:02:09.400 At what is seemingly a botch CrowdStrike update that has 0:02:09.480 --> 0:02:13.400 caused this massive IT outage, as Ed says, disrupting Microsoft 0:02:13.440 --> 0:02:17.360 Global airlines, banks, healthcare systems were hit hard by what 0:02:17.480 --> 0:02:19.400 is said to be one of the few outages of 0:02:19.440 --> 0:02:20.160 this scale. 0:02:20.480 --> 0:02:21.440 It's global in nature. 0:02:21.480 --> 0:02:24.040 Who do we bring on our own Tom Giles for 0:02:24.120 --> 0:02:28.880 the team leading perspective here. I mean, we had a 0:02:28.960 --> 0:02:31.600 very busy set of teams across Asia and Europe today. 0:02:32.000 --> 0:02:33.000 It's great to be here. 0:02:33.040 --> 0:02:38.399 Thanks for having me. Yeah, massive outage, unprecedented in its 0:02:38.400 --> 0:02:43.560 scale in many many ways. As ed so succinctly summarized, 0:02:44.160 --> 0:02:48.360 there was this upgrade to CrowdStrike software. This is a 0:02:48.400 --> 0:02:52.760 cybersecurity team that many companies around the world, many businesses, 0:02:52.800 --> 0:02:58.799 many government agencies rely upon to monitor their systems and 0:02:58.960 --> 0:03:02.800 keep the intrusions out. There was an upgrade and it 0:03:02.880 --> 0:03:07.320 affected our Windows systems around the world and chaos has ensued. 0:03:08.760 --> 0:03:13.119 Okay, so Windows is the operating system design and operated 0:03:13.160 --> 0:03:17.320 by Microsoft. Separately, there was a Microsoft Azure or cloud outage, 0:03:17.440 --> 0:03:20.160 which is confused things. This is the latest from Microsoft 0:03:20.200 --> 0:03:23.680 three sixty five, basically saying that they that it appears 0:03:23.720 --> 0:03:28.919 Microsoft three sixty apps normal services have recovered again, there. 0:03:28.840 --> 0:03:29.720 Is ongoing issues. 0:03:29.760 --> 0:03:30.960 I mean, all you have to do tomm is go 0:03:30.960 --> 0:03:33.040 on social media and see the images of the dreaded 0:03:33.040 --> 0:03:36.880 blue screen that Caroline brought up. What also is interesting 0:03:37.000 --> 0:03:41.400 is is how widespread industries were hit. Airlines, banks, what's 0:03:41.440 --> 0:03:43.360 our coverage telling us about that? 0:03:43.680 --> 0:03:45.640 The global nature of the disruption. 0:03:45.600 --> 0:03:50.240 Banks ups so deliveries are disrupted, flights around the world. 0:03:50.640 --> 0:03:54.800 What this shows is the fragility of our global systems 0:03:54.800 --> 0:04:00.520 that are so reliant upon single vendors and one particular 0:04:00.600 --> 0:04:04.840 patch can just go consent things awry. And when you 0:04:04.880 --> 0:04:08.280 combine that with the ubiquity of Windows machines, they run 0:04:08.640 --> 0:04:11.360 something to the tune of seventy percent of the of 0:04:11.400 --> 0:04:14.640 the computers around the world. And then you, on top 0:04:14.680 --> 0:04:18.599 of that add Azure, this cloud computing system that many companies, 0:04:18.880 --> 0:04:22.960 many businesses around the world rely upon for their cloud systems. 0:04:23.320 --> 0:04:28.719 This really put together a toxic cocktail of it. Of 0:04:28.880 --> 0:04:32.640 it outages everywhere, and it just shows how vulnerable our 0:04:32.680 --> 0:04:35.080 systems are to these single vendors. 0:04:35.640 --> 0:04:39.000 Both Tom Giles, senior executive editor leading our global coverage 0:04:39.000 --> 0:04:40.960 of technology. Later in the show, I promise we're actually 0:04:41.000 --> 0:04:44.680 going to get a little deeper on what endpoint security is. 0:04:44.680 --> 0:04:47.360 We have a brilliant technical guest, but the market reaction, 0:04:47.680 --> 0:04:49.920 let's talk about that and bring in Jordan Klein tech 0:04:50.000 --> 0:04:53.880 media and telecom analysts and Mizuo desk analysts, and my goodness, 0:04:53.920 --> 0:04:55.880 your phone must have been ringing off the hook this morning. 0:04:55.920 --> 0:04:59.160 What have investors basically been telling you about their concern 0:04:59.240 --> 0:04:59.960 or lack of concern? 0:05:00.960 --> 0:05:03.320 Oh good, good afternoon, Ed, Thanks for having me. I 0:05:03.360 --> 0:05:08.760 think clearly people are reacting this way because CrowdStrike is 0:05:08.760 --> 0:05:11.800 probably one of the most owned and liked companies in 0:05:11.839 --> 0:05:15.320 the software sector. It's up one of you know, almost 0:05:15.360 --> 0:05:18.480 thirty five forty percent this year. There's very few companies 0:05:18.480 --> 0:05:21.039 in software that are up that much. And when you 0:05:21.120 --> 0:05:24.239 have one of these events, which thankfully wasn't a hack, 0:05:24.760 --> 0:05:28.680 it was an outage, and there's a huge difference, you know, 0:05:29.000 --> 0:05:30.520 reputationally for the company. 0:05:31.120 --> 0:05:32.000 But when you have. 0:05:32.080 --> 0:05:34.760 Two weeks left in your quarter, they're a July ending 0:05:35.040 --> 0:05:39.479 second quarter, it makes people extra nervous that any disruption 0:05:39.880 --> 0:05:44.400 or problems closing business with new customers to make their estimates. 0:05:44.040 --> 0:05:46.440 Or their guidance could be at risk. 0:05:46.720 --> 0:05:49.920 And right now we're in a position where investors are 0:05:49.960 --> 0:05:53.520 looking for reason to sell tech and rotate into other 0:05:53.680 --> 0:05:56.480 lagging areas, and I think they're kind of using this 0:05:56.600 --> 0:06:00.000 to take risk off and you know, not worry whether 0:06:00.279 --> 0:06:01.320 this impacts their quarter. 0:06:02.200 --> 0:06:04.160 Jordan, I don't know how well us you are in 0:06:04.200 --> 0:06:07.760 the market for endpoint security, but the IDC data shows 0:06:08.240 --> 0:06:11.960 that CrowdStrike has about a fifth of the market. Ironically, 0:06:12.040 --> 0:06:14.719 the number two in that market in terms of individual 0:06:14.800 --> 0:06:16.120 players is Microsoft. 0:06:16.880 --> 0:06:19.840 Both of them under pressure. How do you play that well? 0:06:20.120 --> 0:06:23.680 I think there's it's a good question. I don't really 0:06:23.680 --> 0:06:25.920 think that you play that in the view that one 0:06:26.000 --> 0:06:27.680 is going to lose a bunch of share or they 0:06:27.680 --> 0:06:32.640 both will. Because it's a outage and it's not a hack, 0:06:33.240 --> 0:06:35.320 that to me suggests you're not going to see a 0:06:35.320 --> 0:06:39.200 bunch of customers moving off of CrowdStrike or even Microsoft. 0:06:39.680 --> 0:06:42.599 Some people are, you know, bidding up shares of Sentinel One. 0:06:43.400 --> 0:06:46.919 That's a stock that competes at endpoint with CrowdStrike, but 0:06:47.240 --> 0:06:49.160 I don't even know where they rank in the top 0:06:49.320 --> 0:06:53.480 you know market share listings. Palo Alto is another, you know, 0:06:53.720 --> 0:06:58.160 big competitor. They don't necessarily do all endpoint, but they 0:06:58.160 --> 0:07:03.320 do firewalls and security, and that stock is gaining because people, 0:07:03.400 --> 0:07:06.080 I guess, assume they'll benefit. So at the end of 0:07:06.120 --> 0:07:08.000 the day, we'll have to wait and see what the 0:07:08.320 --> 0:07:13.119 implications or ramifications are. But I would advise investors don't 0:07:13.120 --> 0:07:16.640 make any quick, snap, knee jerk reactions in terms of 0:07:16.800 --> 0:07:17.680 buying or selling. 0:07:18.960 --> 0:07:20.800 And I'm more of a net buyer on. 0:07:20.680 --> 0:07:23.160 This weakness, and as it relates to CrowdStrike, I think 0:07:23.160 --> 0:07:25.680 they'll come through fine in a couple of weeks. People 0:07:25.760 --> 0:07:27.800 won't even be talking about this. 0:07:27.800 --> 0:07:30.239 This is a big market cap impact. 0:07:30.400 --> 0:07:30.720 Jordan. 0:07:31.320 --> 0:07:33.520 What's interesting in your note is you bring up, look, 0:07:33.560 --> 0:07:35.720 are they on the hook financially in any way? 0:07:36.120 --> 0:07:37.440 How will we understand that. 0:07:37.720 --> 0:07:40.880 From CrowdStrike as to whether or not their own clients 0:07:41.360 --> 0:07:45.120 either put off renewing or say, look, I've lost a 0:07:45.120 --> 0:07:47.000 lot of business here, where's the money. 0:07:48.840 --> 0:07:52.080 Yeah, that's a great question. I've talked to some people. 0:07:52.720 --> 0:07:57.440 We put out a quick research update from Mizuho Research Analysts. 0:07:58.880 --> 0:08:00.520 What I think is going to happen is you're not 0:08:00.560 --> 0:08:04.800 going to hear anything until they report their quarter, and 0:08:04.800 --> 0:08:08.440 that won't come until the quarter closes after the end 0:08:08.480 --> 0:08:11.480 of July, so it could be as much as a 0:08:11.520 --> 0:08:14.720 month until you don't really get an official update as 0:08:14.760 --> 0:08:18.200 to what they're liable for. If anything, we do not 0:08:18.400 --> 0:08:20.960 think their contracts are written such that they'll have to 0:08:22.560 --> 0:08:26.360 basically pay these companies for lost business or revenue. But 0:08:26.560 --> 0:08:29.800 what's probably going to happen is to appease companies that 0:08:29.840 --> 0:08:33.720 were disrupted, they will give credits, meaning they will say 0:08:34.520 --> 0:08:37.280 you're under contract and you owe us money, but we 0:08:37.360 --> 0:08:42.360 will give you a credit for business that you owe 0:08:42.400 --> 0:08:46.480 us under the contract. So it's basically a discount. How 0:08:46.559 --> 0:08:49.640 much we don't know, But as soon as they can 0:08:49.720 --> 0:08:52.400 quantify that, I think the markets. 0:08:52.000 --> 0:08:54.920 Will like the clarity and they'll be more. 0:08:54.960 --> 0:08:59.160 Positively disposed once they can quantify what the impact will be. 0:09:00.000 --> 0:09:02.440 Wouldn't imagine it's going to be any massive amount. 0:09:03.600 --> 0:09:06.760 And to that point, when you've got a six billion 0:09:06.840 --> 0:09:09.480 dollar move on a market cap, maybe it's not as 0:09:09.480 --> 0:09:11.720 big as six billion. Maybe you can buy into some 0:09:11.760 --> 0:09:14.680 of this weakness Jordan that you're speaking of. How many 0:09:14.800 --> 0:09:18.640 people do you think, though, in this current environment of jitteriness, 0:09:19.160 --> 0:09:21.600 are willing to go long into a weekend such as this. 0:09:22.240 --> 0:09:24.400 We're going to find out as the market, you know, 0:09:25.040 --> 0:09:26.920 we go into the afternoon and we see how this 0:09:27.000 --> 0:09:30.800 stock trades, I would say, initially, you know, it's eleven 0:09:30.800 --> 0:09:34.720 o'clock Eastern, I'm pretty encouraged by the way the stock's acting, 0:09:35.320 --> 0:09:36.920 given it's well off the lows. 0:09:37.200 --> 0:09:38.400 If this was something. 0:09:38.120 --> 0:09:41.800 Where people were in an absolute panic and in the 0:09:41.840 --> 0:09:44.440 market was melting down, I think it would be at 0:09:44.480 --> 0:09:47.800 the lows, and it's not, so I wouldn't be surprised 0:09:47.840 --> 0:09:52.160 if it continues to grind a bit higher. But I 0:09:52.200 --> 0:09:56.480 do think again, the sentiment and tech broadly is people 0:09:56.679 --> 0:10:00.400 are nervous, the sector is not acting well. It's it's 0:10:00.559 --> 0:10:04.280 particularly you know, nerve racking for those as we wait 0:10:04.320 --> 0:10:07.360 for earnings, and I kind of feel like most investors 0:10:07.440 --> 0:10:09.120 are going to take a wait and see. You know, 0:10:09.160 --> 0:10:12.760 our volumes on our trading desk don't suggest panic selling, 0:10:12.800 --> 0:10:15.640 but it doesn't suggest you're seeing waves of buyers coming 0:10:15.679 --> 0:10:18.120 in after a few days of weakness either. 0:10:18.679 --> 0:10:21.240 Jordan Klein, perfect voice for today. We thank you tech 0:10:21.320 --> 0:10:23.400 media and telecom analyst at Mizuho. 0:10:23.679 --> 0:10:24.680 Appreciate the time. 0:10:25.280 --> 0:10:28.280 Now let's just talk about Meta CEO and founder Mark 0:10:28.360 --> 0:10:31.560 Zuckerberg wing in on the current election landscape in an 0:10:31.600 --> 0:10:34.080 interview with Bluembgo Original's host Emily Chang for an episode 0:10:34.120 --> 0:10:35.480 of the Circuit Just take a listen. 0:10:36.320 --> 0:10:39.079 The main thing that I hear from people is that 0:10:39.720 --> 0:10:41.920 they actually want to see less political content on our 0:10:41.960 --> 0:10:44.760 services because they come to our services to connect with people. 0:10:45.320 --> 0:10:48.800 So you know, that's what we're going to do where 0:10:49.280 --> 0:10:52.000 we give people control over this, but we're generally trying 0:10:52.040 --> 0:10:55.440 to recommend less political content. So I think you're going 0:10:55.480 --> 0:10:57.360 to see our services play less of a role in 0:10:57.400 --> 0:10:58.800 this election than they have in the past. 0:10:59.160 --> 0:11:01.760 What's the full use of interview with Meta CEO Mark 0:11:01.800 --> 0:11:03.520 Zuckerberg on the circuit with Emily Chang. 0:11:03.720 --> 0:11:05.520 It's July twenty third, six thirty. 0:11:05.280 --> 0:11:08.360 Pm in New York on Bloomberg Television or stream it 0:11:08.400 --> 0:11:18.080 eight thirty pm ET on Bloomberg Originals. 0:11:20.040 --> 0:11:22.200 In terms of that member growth insurre, I'd say that 0:11:22.559 --> 0:11:25.559 the kind of outside paid outsized paidin and ads in 0:11:25.600 --> 0:11:28.520 the quarter was primarily driven by a stronger acquisition, a 0:11:28.559 --> 0:11:31.800 little stronger than we expected, but also very healthy, continued 0:11:31.840 --> 0:11:34.840 healthy retention in the quarter. And that's across all regions. 0:11:37.000 --> 0:11:38.600 That was the Netflix CFO. 0:11:38.720 --> 0:11:41.840 They're talking about the company's member or subscriber growth for 0:11:41.880 --> 0:11:46.080 retention earning showed strong subscriber editions in the second quarter. 0:11:46.280 --> 0:11:48.640 I want to bring a Needham senior Entertainment Internet analyst, 0:11:48.679 --> 0:11:52.720 Laura Martin. It's a situation where after hours stock go down, 0:11:53.320 --> 0:11:58.600 market open stock go up. Good subscriber numbers. Concerning guidance, 0:11:58.960 --> 0:12:00.439 which is more important for you, Laura. 0:12:02.200 --> 0:12:05.239 I think the subscriber number is in focus in both periods, 0:12:05.280 --> 0:12:08.240 so they over delivered. They added eight million subscribers that 0:12:08.280 --> 0:12:10.760 was about twice the consensus. But then they had a 0:12:10.880 --> 0:12:13.720 very anemic subscriber growth number for the third quarter, and 0:12:13.760 --> 0:12:17.000 they talked about a tougher second half. So the answer 0:12:17.080 --> 0:12:21.120 is subscribers matter a lot, but not that subscribers like overall. 0:12:21.360 --> 0:12:24.600 And also the location of the subscribers is sort of suboptimal. 0:12:24.920 --> 0:12:28.480 They're adding subscribers in the lowest RPOO or what they 0:12:28.480 --> 0:12:31.720 call arm average revenue per user, and so that's actually 0:12:31.760 --> 0:12:34.079 Wall Street doesn't like that either. And the biggest, biggest 0:12:34.080 --> 0:12:37.400 problem is they're going to stop giving us subscriber guidance 0:12:37.840 --> 0:12:40.839 and actually our POO guidance in the first quarter of 0:12:40.880 --> 0:12:42.720 next year, and they just want to talk about revenue. 0:12:42.800 --> 0:12:44.520 And that's a really big problem because it's going to 0:12:44.559 --> 0:12:45.959 add volatility to the stock. 0:12:47.120 --> 0:12:49.360 Where the subscribers are I find so interesting as well. 0:12:49.360 --> 0:12:51.839 So they talked about strength in Asia Pacific and then 0:12:52.320 --> 0:12:55.000 in terms of context of the guidance, like currency impacts, 0:12:55.040 --> 0:12:58.960 particularly in Argentina. If you're covering this stock, where do 0:12:59.040 --> 0:13:02.200 you want to see subscribe a growth sixteen. 0:13:01.840 --> 0:13:04.160 Dollars artpoo's in America? You want to see them all 0:13:04.200 --> 0:13:07.720 there because it's like two p's they pay twice as 0:13:07.760 --> 0:13:10.480 much in America compared to latam or apath. 0:13:10.880 --> 0:13:15.160 Okay, see you one more expensive money being taken in 0:13:15.280 --> 0:13:17.920 from the more expensive clientele. But ultimately they're going to 0:13:17.920 --> 0:13:20.400 get out of the Monaaura. We saturated. We're still trying 0:13:20.440 --> 0:13:22.920 to descend which streaming service I want to be using 0:13:22.920 --> 0:13:23.960 and which one should be merging. 0:13:24.400 --> 0:13:24.640 Yep. 0:13:24.840 --> 0:13:27.080 Well, and I really like their going. I think the 0:13:27.120 --> 0:13:30.920 reason their churn is falling, like in America they're saturated, 0:13:30.960 --> 0:13:33.520 as you say, is live. I think when they do 0:13:33.559 --> 0:13:36.040 these live fights and these live roasts and these like 0:13:36.200 --> 0:13:39.640 exclusive live programming, it keeps people paying for the next 0:13:39.679 --> 0:13:44.079 month because they really market these things, these exclusive live events. 0:13:44.280 --> 0:13:47.079 So I really like, like, I think their video game 0:13:47.360 --> 0:13:51.640 and advertising is very troubled. Video games troubled, but I 0:13:51.679 --> 0:13:54.280 think live is a really interesting genre for Netflix to 0:13:54.320 --> 0:13:54.839 lower turn. 0:13:55.200 --> 0:13:57.760 Okay, so they need to be investing basically because that 0:13:57.840 --> 0:14:01.200 takes a herculean effort from a technology perspective, right. 0:14:01.160 --> 0:14:03.320 Yeah, And they're paying for like NFL games, so they're 0:14:03.360 --> 0:14:08.439 joining the arms race in really expensive live sports programming. 0:14:08.480 --> 0:14:11.160 So that isn't great for your content budget, you know. 0:14:11.200 --> 0:14:12.480 Look ahead, Laura. 0:14:12.800 --> 0:14:15.199 We could go on and on about this entire environment, 0:14:15.240 --> 0:14:17.440 and we didn't even get to some of the discussion 0:14:17.480 --> 0:14:20.120 around the M and A today as well. Ninham Senior 0:14:20.200 --> 0:14:23.480 Entertainment and Internet analyst Laura Martin always so spot on 0:14:23.520 --> 0:14:25.640 when it comes to this market analysis. Meanwhile, and we 0:14:25.680 --> 0:14:27.720 just want to turn our attention to another earnings today 0:14:27.920 --> 0:14:30.320 when we would say at a large fintech American Express 0:14:30.400 --> 0:14:33.240 at the stock trading well, it's been a bumpy ride today, 0:14:33.280 --> 0:14:35.640 in fact, currently off by more than four percent as 0:14:35.680 --> 0:14:38.240 the company sees fiscal year earnings for share between thirteen 0:14:38.360 --> 0:14:41.640 thirty and thirteen eighty dollars, and it actually was looking 0:14:41.680 --> 0:14:44.160 pretty strong, but it's warning that it plans to increase 0:14:44.280 --> 0:14:47.000 spending on marketing in the coming months. CEO sc Scurry 0:14:47.040 --> 0:14:49.560 telling me earlier today that they're really committed to marketing 0:14:49.880 --> 0:14:53.960 to tech investments, particularly in artificial intelligence at the moment 0:14:54.120 --> 0:14:57.640 that this is a company that has had record revenue overall, 0:14:57.920 --> 0:15:01.320 but the market's still questioning that's how they continue to 0:15:01.360 --> 0:15:04.080 be beefing up on the profitability side. Like many other 0:15:04.120 --> 0:15:06.360 American spects, has also been seeing an impact to the 0:15:06.400 --> 0:15:08.440 crowd strike outage, saying quote, we. 0:15:08.400 --> 0:15:09.720 Had to slow down early on. 0:15:09.840 --> 0:15:12.200 All airports were impacted, so our call logs were very 0:15:12.280 --> 0:15:16.640 high as people tried to change their flights, so CEO 0:15:16.800 --> 0:15:19.840 telling us that at the moment, though still pretty confident. 0:15:19.960 --> 0:15:20.920 Ed what we got coming up. 0:15:21.160 --> 0:15:23.800 Another big name in the world of software, Salesforce announced 0:15:23.880 --> 0:15:27.280 launch of the Einstein Service Agent, their first fully autonomous 0:15:27.320 --> 0:15:32.240 AI agent for customer support. Salesforce AI CEO Clarashi joining 0:15:32.280 --> 0:15:32.720 us next. 0:15:32.760 --> 0:15:33.920 This is Bluembo Technology. 0:15:48.800 --> 0:15:50.240 Okay, it's the growing field of AI. 0:15:50.480 --> 0:15:54.640 Salesforce recently announcing the launch of their Einstein Service Agent. 0:15:55.000 --> 0:15:57.720 This is the company's first fully autonomous AI agent working 0:15:57.760 --> 0:16:01.880 to transform customer service jobs. That's bringing Clarashi, CEO of 0:16:01.920 --> 0:16:07.280 Salesforce AI for more. Clara, how can you quantify the 0:16:07.320 --> 0:16:09.480 difference of a normal, typical chatbot here? 0:16:09.520 --> 0:16:11.840 How much better is this agent you're offering? 0:16:12.040 --> 0:16:14.160 Thank you so much for having me. You know, we've 0:16:14.200 --> 0:16:17.880 all interacted with customer support chatbots and what they can 0:16:17.920 --> 0:16:20.800 do is pretty limited. It's it's always obvious that you're 0:16:20.840 --> 0:16:24.120 talking with a bot. Einstein Service Agent is our new 0:16:24.160 --> 0:16:29.080 offering that completely redefines the self service experience. Really feels 0:16:29.080 --> 0:16:32.080 like you're chatting with a human expert who can answer 0:16:32.480 --> 0:16:36.880 complex questions and take complicated actions, you know, for example, 0:16:37.240 --> 0:16:40.560 looking up in order, reviewing a photo that you've chatted 0:16:40.600 --> 0:16:44.760 in of a damaged product, issuing a return shipping label, 0:16:45.120 --> 0:16:49.200 sending a replacement multi steps, all in a secure and 0:16:49.240 --> 0:16:53.120 trusted way. Because Einstein is grounded in your company's trusted 0:16:53.200 --> 0:16:56.520 data policies and procedures in your salesforce CRM. 0:16:57.120 --> 0:17:01.120 There's a multi modal element to this. How much have 0:17:01.280 --> 0:17:05.240 you been building the large language models yourself depending on others? 0:17:05.480 --> 0:17:07.359 Where does ultimately this agent come from? 0:17:07.560 --> 0:17:10.320 Well, we've been developing large language models since twenty eighteen. 0:17:10.359 --> 0:17:13.680 We acquired a company called metamind and our Salesforce research group. 0:17:13.920 --> 0:17:18.400 They are one of the most prolific publishers of Frontier Research. 0:17:18.680 --> 0:17:22.520 That being said, we believe in an open architecture because 0:17:22.560 --> 0:17:25.280 we want to give customers choice when it comes to 0:17:25.400 --> 0:17:29.360 AI models, and we've recently launched our CRM leaderboard. It's 0:17:29.400 --> 0:17:35.120 the first LM leaderboard specifically for sales, service, marketing and commerce. 0:17:35.320 --> 0:17:38.639 And we found something amazing, which is first that small models, 0:17:38.680 --> 0:17:43.360 in some cases open source models actually outperform the biggest, 0:17:43.640 --> 0:17:46.199 most expensive models on the market when it comes to 0:17:46.240 --> 0:17:49.200 a lot of enterprise tasks. The second thing that we 0:17:49.320 --> 0:17:53.240 learned is that the leaderboard is constantly changing week over weekends. 0:17:53.240 --> 0:17:56.280 So we really feel as important to support an open, 0:17:56.400 --> 0:17:57.919 vibrant ecosystem of models. 0:17:58.280 --> 0:18:00.399 It's Clark, good morning, It's Ed in Cisco. 0:18:00.440 --> 0:18:04.160 Look at our lead story is the global outage CrowdStrike Microsoft. 0:18:04.400 --> 0:18:08.440 There's some three sixty five integration with different Salesforce platforms 0:18:08.480 --> 0:18:10.919 and offerings, and you know your role across AI but 0:18:10.960 --> 0:18:14.040 also cloud. What has the impact been on you and 0:18:14.119 --> 0:18:16.040 Salesforce and what's the response been. 0:18:16.080 --> 0:18:18.920 Please, I have a number of friends who are stuck 0:18:18.960 --> 0:18:20.800 in the airport right now. You know, I think nearly 0:18:20.840 --> 0:18:24.119 every industry around the world has been affected. I'm Salesforce 0:18:24.359 --> 0:18:27.879 included by the crowd strike issue. Our engineers are working 0:18:28.080 --> 0:18:30.760 with the CD strike team around the clock to restore 0:18:30.800 --> 0:18:32.240 full access to our services. 0:18:33.200 --> 0:18:35.760 And just finally on that topic, you know you've long 0:18:35.800 --> 0:18:39.040 been in technology. How does this event rank, you know, 0:18:39.240 --> 0:18:43.240 in sort of technological disruptional outages you've seen in your career. 0:18:43.680 --> 0:18:46.320 You know, I think there's you know, that's inevitability when 0:18:46.359 --> 0:18:49.119 it comes to technology, and it really underscores the importance 0:18:49.520 --> 0:18:53.240 of working with trusted partners. I mean, trust is Salesforce's 0:18:53.320 --> 0:18:56.159 number one value, and we very much pride ourselves in 0:18:56.160 --> 0:19:00.320 our uptime partnering with many customers, specifically an AI you know, 0:19:00.400 --> 0:19:04.280 Triple A Air India Open Table. These are companies that 0:19:04.359 --> 0:19:08.360 really help run the world, and so we take that trust, security, reliability, 0:19:08.359 --> 0:19:09.640 and performance very seriously. 0:19:10.280 --> 0:19:11.919 Power there's going to be quite a few friends of 0:19:11.960 --> 0:19:14.600 yours at airports and many others who are desperately trying 0:19:14.600 --> 0:19:17.199 to get hold of a real person right now on 0:19:17.240 --> 0:19:21.120 the phone rather than a chatbot, and I'm wondering how 0:19:21.160 --> 0:19:24.280 long that will be inevitable You in many ways do 0:19:24.400 --> 0:19:27.840 transfer conversations with this new AI agent to a human 0:19:28.040 --> 0:19:32.159 if perhaps just gets out of the capability of the agent. 0:19:32.280 --> 0:19:33.920 How often is that likely to happen? Do you think? 0:19:34.960 --> 0:19:36.760 Well, I think that it really depends on, you know, 0:19:36.800 --> 0:19:40.399 the nature of how much knowledge and trusted policies and 0:19:40.440 --> 0:19:44.800 procedures the companies have been documented inside of their organization, 0:19:44.840 --> 0:19:46.480 and if they don't have very much, we're going to 0:19:46.480 --> 0:19:50.399 help them with that AI generated AI drafted knowledge creation. 0:19:50.560 --> 0:19:52.120 But I think by and large we're going to see 0:19:52.160 --> 0:19:54.919 more and more, just like when you're talking to a 0:19:55.040 --> 0:19:59.080 chatbot or talking to chat GPT, sometimes it's hard to 0:19:59.119 --> 0:20:01.879 tell that it's not a human. Now, the challenge is 0:20:01.920 --> 0:20:06.360 that without that contextual data and without those guardrails, these 0:20:06.520 --> 0:20:09.119 llms can just make things up, they can hallucinate. And 0:20:09.160 --> 0:20:12.840 so that's really where Salesforce is Einstein trust layer. That's 0:20:12.880 --> 0:20:17.040 the layer that every single AI interaction goes through before 0:20:17.080 --> 0:20:19.879 it gets presented to a customer. That's so important to 0:20:19.920 --> 0:20:22.520 make sure that it's trusted, reliable, and accurate. 0:20:22.920 --> 0:20:26.040 Clara, our audience has kind of followed the rollout of 0:20:26.040 --> 0:20:29.520 what Einstein is doing and what Salesforce has done in 0:20:29.560 --> 0:20:32.080 the domain of sort of generative AI tool quite closely. 0:20:32.920 --> 0:20:37.119 You must benchmark against other similar platforms or products. What 0:20:37.160 --> 0:20:41.760 have you learned about a performance essentially against what else 0:20:41.840 --> 0:20:42.400 is out there? 0:20:42.800 --> 0:20:44.640 Well, it's such a great question. It's something that we're 0:20:44.720 --> 0:20:49.600 constantly benchmarking. I mentioned our CRM Large Language Model leaderboard 0:20:49.640 --> 0:20:53.560 earlier and we found that there's really three critical ingredients 0:20:53.600 --> 0:20:56.959 to success when it comes to enterprise AI. First and 0:20:57.000 --> 0:21:00.439 foremost is trust. Making sure that you have data privacy, 0:21:00.800 --> 0:21:04.080 data security, ethical guardrails built in. Again, this is our 0:21:04.080 --> 0:21:09.520 Einstein trust layer. Everything from citations to audit trails, to observability, 0:21:09.560 --> 0:21:13.800 monitoring and data masking. The second ingredient is having the 0:21:13.920 --> 0:21:17.840 right context. If you don't give the LLM enough context, 0:21:18.080 --> 0:21:21.560 it'll just make things up. Now you think about our customers. 0:21:21.600 --> 0:21:26.680 They've been putting their trusted data and their trusted business processes, 0:21:26.880 --> 0:21:30.639 policies and procedures into Salesforce for the last twenty five years. 0:21:31.000 --> 0:21:33.520 And with the introduction of our data cloud product now 0:21:33.520 --> 0:21:37.200 where we have that trusted data across the entire enterprise 0:21:37.800 --> 0:21:40.040 of siloed systems, we're bringing that together and so that 0:21:40.160 --> 0:21:45.399 context is critical to accuracy, reliability, and really the usefulness 0:21:45.400 --> 0:21:48.360 of the AI. And then the third ingredient is empowerment 0:21:48.600 --> 0:21:52.520 and helping bring every employee at an organization along for 0:21:52.600 --> 0:21:54.800 this AI learning journey which is going to be which 0:21:54.840 --> 0:21:57.080 is never ending. And I'm really proud of the work 0:21:57.119 --> 0:22:00.320 that we've done with Trailhead. We've open sourced all of 0:22:00.359 --> 0:22:03.720 the courses that we developed for our internal employees Clark 0:22:03.840 --> 0:22:04.919