1 00:00:04,680 --> 00:00:12,840 Speaker 1: Welcome to Tech Stuff, a production from iHeartRadio. Today, we 2 00:00:12,880 --> 00:00:15,880 Speaker 1: are witnessed to one of those rare moments in history, 3 00:00:16,239 --> 00:00:19,439 Speaker 1: the rise of an innovative technology with the potential to 4 00:00:19,600 --> 00:00:24,320 Speaker 1: radically transform business and society forever. That technology, of course, 5 00:00:24,800 --> 00:00:28,360 Speaker 1: is artificial intelligence, and it's the central focus for this 6 00:00:28,520 --> 00:00:32,519 Speaker 1: new season of Smart Talks with IBM. Join hosts from 7 00:00:32,560 --> 00:00:36,280 Speaker 1: your favorite Pushkin podcasts as they talk with industry experts 8 00:00:36,280 --> 00:00:39,920 Speaker 1: and leaders to explore how businesses can integrate AI into 9 00:00:39,960 --> 00:00:43,280 Speaker 1: their workflows and help drive real change in this new 10 00:00:43,360 --> 00:00:47,040 Speaker 1: era of AI, and of course, host Malcolm Gladwell will 11 00:00:47,040 --> 00:00:49,360 Speaker 1: be there to guide you through the season and throw 12 00:00:49,440 --> 00:00:52,360 Speaker 1: in his two cents as well. Look out for new 13 00:00:52,360 --> 00:00:55,280 Speaker 1: episodes of Smart Talks with IBM every other week on 14 00:00:55,320 --> 00:00:59,560 Speaker 1: the iHeartRadio app, Apple Podcasts, wherever you get your podcasts. 15 00:01:00,080 --> 00:01:04,000 Speaker 1: Learn more at IBM dot com slash smart Talks. 16 00:01:06,319 --> 00:01:09,920 Speaker 2: Hello, Hello, Welcome to Smart Talks with IBM, a podcast 17 00:01:09,959 --> 00:01:16,319 Speaker 2: from Pushkin Industries, iHeartRadio and IBM. I'm Malcolm Glappo. This season, 18 00:01:16,360 --> 00:01:20,479 Speaker 2: we're continuing our conversations with new creators visionaries who are 19 00:01:20,560 --> 00:01:25,280 Speaker 2: creatively applying technology and business to drive change, but with 20 00:01:25,319 --> 00:01:30,080 Speaker 2: a focus on the transformative power of artificial intelligence and 21 00:01:30,120 --> 00:01:33,320 Speaker 2: what it means to leverage AI as a game changing 22 00:01:33,440 --> 00:01:39,080 Speaker 2: multiplier for your business. Today's episode highlights the power of collaboration. 23 00:01:39,880 --> 00:01:42,600 Speaker 2: IBM has long been a supporter of the better Together 24 00:01:42,720 --> 00:01:47,000 Speaker 2: mindset and embrace his partnerships. They have been working together 25 00:01:47,040 --> 00:01:50,920 Speaker 2: with Salesforce for more than two decades, but have recently 26 00:01:51,000 --> 00:01:56,600 Speaker 2: launched a new collaborative effort surrounding generative AI. Pushkin's very 27 00:01:56,600 --> 00:02:00,800 Speaker 2: own Jacob Goldstein sat down with Matt Candy and sus Emerson. 28 00:02:01,360 --> 00:02:04,920 Speaker 2: Matt is the global managing Partner of Generative AI at 29 00:02:04,960 --> 00:02:09,360 Speaker 2: IBM Consulting, helping clients and partners around the world embrace 30 00:02:09,440 --> 00:02:14,000 Speaker 2: this new era of technology, and Susan is a senior 31 00:02:14,080 --> 00:02:19,400 Speaker 2: vice president for Salesforce dedicated to AI, analytics and data. 32 00:02:20,280 --> 00:02:23,880 Speaker 2: They discussed the historic collaboration between the two tech giants, 33 00:02:24,480 --> 00:02:28,560 Speaker 2: explored the opportunity AI presents for customer service, and walk 34 00:02:28,639 --> 00:02:35,480 Speaker 2: through how businesses can use generative AI to interface with clients. Okay, 35 00:02:35,800 --> 00:02:37,080 Speaker 2: let's get to the conversation. 36 00:02:39,680 --> 00:02:42,919 Speaker 3: Thank you guys for coming this morning. So I'm interested 37 00:02:42,960 --> 00:02:46,760 Speaker 3: in how you both came to generative AI, or maybe 38 00:02:46,760 --> 00:02:48,120 Speaker 3: it sort of came to you in the way it 39 00:02:48,160 --> 00:02:50,160 Speaker 3: sort of came to all of us, But how did 40 00:02:50,200 --> 00:02:52,360 Speaker 3: you arrive at working on generative AI. 41 00:02:52,760 --> 00:02:55,560 Speaker 4: As part of my remitted Salesforce. Over the years, I've 42 00:02:55,600 --> 00:02:59,280 Speaker 4: brought a lot of analytics and data and machine learning 43 00:02:59,280 --> 00:03:04,040 Speaker 4: products to life under the Einstein brand at Salesforce. So 44 00:03:04,520 --> 00:03:09,480 Speaker 4: as we pivoted Salesforce into taking advantage of the generative 45 00:03:09,480 --> 00:03:12,600 Speaker 4: AI moment, it was natural that I became part of 46 00:03:12,639 --> 00:03:19,519 Speaker 4: the advanced team leveraging generative AI, and it's become interesting. 47 00:03:19,639 --> 00:03:23,360 Speaker 4: But what I see as I speak with customers the 48 00:03:23,440 --> 00:03:25,840 Speaker 4: moment that everyone is facing in terms of how they 49 00:03:26,040 --> 00:03:31,959 Speaker 4: incorporate genitive AI into their businesses, their workforces, and their 50 00:03:32,000 --> 00:03:35,280 Speaker 4: technical stacks. It's actually opening up a lot of doors 51 00:03:35,360 --> 00:03:40,320 Speaker 4: to other utility of analytics, data and AI. So it's 52 00:03:40,360 --> 00:03:44,400 Speaker 4: been this big pull through in terms of incorporating not 53 00:03:44,520 --> 00:03:48,640 Speaker 4: just generative AI, but a larger conversation around how we 54 00:03:49,080 --> 00:03:52,720 Speaker 4: become all better using data in our day jobs. 55 00:03:54,280 --> 00:03:56,960 Speaker 3: So that's a great frame for sort of what's going 56 00:03:57,000 --> 00:04:00,840 Speaker 3: on at Salesforce with generative AI. Tell us a little 57 00:04:00,880 --> 00:04:03,080 Speaker 3: bit about you know, how that fits with the way 58 00:04:03,520 --> 00:04:04,920 Speaker 3: IBM is approaching with space. 59 00:04:05,680 --> 00:04:08,880 Speaker 5: Yeah, so I guess through three sides to that question. 60 00:04:09,520 --> 00:04:12,360 Speaker 5: And so there's the technology side of it. So IBM 61 00:04:12,440 --> 00:04:15,720 Speaker 5: has a technology organization, and so you know, we are 62 00:04:15,760 --> 00:04:19,400 Speaker 5: building and have been over many years decades. In fact, 63 00:04:19,480 --> 00:04:22,919 Speaker 5: IBM has been working in this space a generative AI 64 00:04:23,080 --> 00:04:30,600 Speaker 5: stack that allows organizations to adopt generative AI technology aimed 65 00:04:30,640 --> 00:04:35,200 Speaker 5: at enterprise and business use within their organizations. So then 66 00:04:35,240 --> 00:04:37,839 Speaker 5: within the consulting business, you know, we have one hundred 67 00:04:37,839 --> 00:04:41,400 Speaker 5: and sixty thousand people who work every day with clients 68 00:04:41,440 --> 00:04:47,080 Speaker 5: across every industry, regulated industries, government organizations, and so this, 69 00:04:47,320 --> 00:04:50,800 Speaker 5: you know, is a really important technology that those companies 70 00:04:50,839 --> 00:04:53,080 Speaker 5: are going to be using to drive the next level 71 00:04:53,080 --> 00:04:57,359 Speaker 5: of transformation in their enterprises processes and the types of 72 00:04:57,400 --> 00:05:00,479 Speaker 5: experiences they build for their customers. And so you know, 73 00:05:00,520 --> 00:05:05,880 Speaker 5: we work extensively with partners technology such as Salesforce, AWS, Microsoft, 74 00:05:06,400 --> 00:05:09,760 Speaker 5: as well as our own technology. And then finally, I 75 00:05:09,760 --> 00:05:12,600 Speaker 5: guess the third angle is the work that we've got 76 00:05:12,640 --> 00:05:15,560 Speaker 5: to do to reinvent the business of consulting. And so 77 00:05:15,720 --> 00:05:18,880 Speaker 5: if I think about you know, consulting in systems integration, 78 00:05:19,320 --> 00:05:22,839 Speaker 5: you know, ultimately we are knowledge workers, right, and so 79 00:05:23,120 --> 00:05:25,880 Speaker 5: from an industry perspective, I think, you know, our industry 80 00:05:25,960 --> 00:05:29,080 Speaker 5: is same as many others is it's going to go 81 00:05:29,279 --> 00:05:33,359 Speaker 5: undergo a level of disruption caused by this technology, but 82 00:05:33,440 --> 00:05:36,160 Speaker 5: therefore that will also create a huge opportunity for us 83 00:05:36,640 --> 00:05:37,039 Speaker 5: as well. 84 00:05:37,120 --> 00:05:40,080 Speaker 6: So those three aspects, Jacob, great. 85 00:05:39,920 --> 00:05:42,040 Speaker 3: So, so that's the point of view sort of from 86 00:05:42,080 --> 00:05:46,279 Speaker 3: your companies in your work. I'm curious to talk for 87 00:05:46,360 --> 00:05:48,960 Speaker 3: a moment about AI from the point of view of 88 00:05:49,720 --> 00:05:53,599 Speaker 3: consumers and employees kind of out in the world today. 89 00:05:53,640 --> 00:05:56,560 Speaker 3: So just to start with consumers, when I'm just out 90 00:05:57,160 --> 00:05:59,479 Speaker 3: as a person as a consumer in the world, how 91 00:05:59,480 --> 00:06:01,320 Speaker 3: am I AI today? 92 00:06:02,480 --> 00:06:03,840 Speaker 6: I'll give you a great little use case. 93 00:06:03,880 --> 00:06:07,919 Speaker 5: Actually, I was on holiday three weeks ago in Tenerif 94 00:06:07,960 --> 00:06:10,919 Speaker 5: in Spain, and I was trying to find somewhere to 95 00:06:10,960 --> 00:06:11,960 Speaker 5: park the car with the. 96 00:06:11,920 --> 00:06:13,360 Speaker 6: Family for dinner that evening. 97 00:06:14,000 --> 00:06:19,240 Speaker 5: And I found this area next to this kind of 98 00:06:19,279 --> 00:06:22,359 Speaker 5: shopping center and there was this sign there and I 99 00:06:22,360 --> 00:06:24,719 Speaker 5: couldn't quite work out if it was saying I could 100 00:06:24,760 --> 00:06:27,760 Speaker 5: park there or not, And so I took a photo 101 00:06:27,839 --> 00:06:30,719 Speaker 5: of the sign and I uploaded it to an AI 102 00:06:31,080 --> 00:06:33,279 Speaker 5: tool and I said, what does this mean? And it 103 00:06:33,320 --> 00:06:35,479 Speaker 5: basically explained to me what the sign was saying and 104 00:06:35,520 --> 00:06:37,640 Speaker 5: basically told me that I shouldn't be parking there, and 105 00:06:37,680 --> 00:06:40,920 Speaker 5: so I drove on and I found some somewhere else 106 00:06:40,960 --> 00:06:45,520 Speaker 5: to park. But you know, that allowed me in under 107 00:06:45,560 --> 00:06:49,440 Speaker 5: sixty seconds to probably avoid one hundred euro fine by 108 00:06:49,480 --> 00:06:52,960 Speaker 5: parking the car there. So just a simple example, but 109 00:06:53,080 --> 00:06:56,400 Speaker 5: I think the ability that these tools have to take 110 00:06:56,480 --> 00:06:59,320 Speaker 5: friction out of our daily lives, you know, and to 111 00:06:59,400 --> 00:07:02,400 Speaker 5: be able to make just things that we do in 112 00:07:02,440 --> 00:07:06,040 Speaker 5: our everyday life simple and more frictionless. You know. That's 113 00:07:06,080 --> 00:07:09,080 Speaker 5: how I look at how mat the consumer is going 114 00:07:09,120 --> 00:07:11,120 Speaker 5: to benefit from some of this type of technology. 115 00:07:11,840 --> 00:07:15,400 Speaker 4: And from my perspective, it's also a travel story. I 116 00:07:15,440 --> 00:07:18,480 Speaker 4: spend a lot of time on the road for work, 117 00:07:18,840 --> 00:07:22,840 Speaker 4: but recently had to send my sister and her family 118 00:07:23,000 --> 00:07:25,960 Speaker 4: to a destination they had never been to for a wedding. 119 00:07:26,640 --> 00:07:30,240 Speaker 4: And it was really quick and easy to use some 120 00:07:30,600 --> 00:07:33,520 Speaker 4: generitive tools to come up with a whole plan for 121 00:07:33,560 --> 00:07:35,800 Speaker 4: them because they love to hike and to be outdoors 122 00:07:36,160 --> 00:07:39,920 Speaker 4: and to hike in areas that aren't overly crowded with people, 123 00:07:40,520 --> 00:07:43,840 Speaker 4: and so Jenai very quickly gave me an itinerary of 124 00:07:43,880 --> 00:07:47,280 Speaker 4: a bunch of terrific hikes for them for a destination. 125 00:07:47,680 --> 00:07:48,720 Speaker 4: So things like that. 126 00:07:48,920 --> 00:07:53,440 Speaker 3: Great, And then what about the effect of AI and 127 00:07:53,480 --> 00:07:57,600 Speaker 3: of automation more generally on employees on the workforce. 128 00:07:58,280 --> 00:08:01,360 Speaker 4: Well, there's so many dimensions to take that from. Generative 129 00:08:01,400 --> 00:08:05,119 Speaker 4: AI really can up level a workforce in all sorts 130 00:08:05,120 --> 00:08:09,400 Speaker 4: of ways by providing these consistent ways to engage with technology, 131 00:08:09,480 --> 00:08:13,080 Speaker 4: with these natural language experiences. So I think it changes 132 00:08:13,120 --> 00:08:17,200 Speaker 4: everything from it finds us content, it generates us content, 133 00:08:17,480 --> 00:08:20,200 Speaker 4: It makes it easier to work with our systems of 134 00:08:20,240 --> 00:08:25,520 Speaker 4: engagement and operation, and for many organizations it can be 135 00:08:25,960 --> 00:08:29,120 Speaker 4: a lifting factor in terms of bringing a more consistent 136 00:08:29,400 --> 00:08:33,600 Speaker 4: workforce experience because these tools can just be ever present 137 00:08:33,840 --> 00:08:35,800 Speaker 4: in our systems of work. 138 00:08:36,679 --> 00:08:38,760 Speaker 5: I mean, I'll give you a little example here in IBM, 139 00:08:38,880 --> 00:08:42,240 Speaker 5: we have something called our Skajar and so that's our 140 00:08:42,320 --> 00:08:47,320 Speaker 5: conversational AI interface that we use to interact with HR 141 00:08:47,520 --> 00:08:52,200 Speaker 5: services and ninety four percent of every employee interaction now 142 00:08:52,640 --> 00:08:56,040 Speaker 5: happens without human intervention through that interface. 143 00:08:56,360 --> 00:08:57,320 Speaker 6: But you would never know that. 144 00:08:58,280 --> 00:09:01,560 Speaker 5: And so if I think about, you know, our HR processes, 145 00:09:01,760 --> 00:09:05,480 Speaker 5: You know, we have this amazing conversational based AI that 146 00:09:05,520 --> 00:09:08,520 Speaker 5: we use for all of our HR interactions, and we 147 00:09:08,640 --> 00:09:12,199 Speaker 5: surface that through SLACK and so Slack becomes the front 148 00:09:12,240 --> 00:09:15,200 Speaker 5: door for how we access a lot of these different 149 00:09:15,320 --> 00:09:19,600 Speaker 5: enterprise processes and capabilities and how we surface AI. In fact, 150 00:09:19,600 --> 00:09:21,640 Speaker 5: I'm taking a flight shortly back to the UK and 151 00:09:22,120 --> 00:09:24,600 Speaker 5: our our skar Bos is reminding me that it's raining 152 00:09:24,600 --> 00:09:26,640 Speaker 5: in the UK and I should take an umbrella. 153 00:09:27,440 --> 00:09:29,880 Speaker 4: Isn't it always like raining in England? 154 00:09:32,240 --> 00:09:34,320 Speaker 5: Yeah, I don't think there's any AI needed for that. 155 00:09:34,360 --> 00:09:37,720 Speaker 5: I think that's just a hard coded If England, then 156 00:09:37,760 --> 00:09:38,480 Speaker 5: take umbrella. 157 00:09:38,520 --> 00:09:39,640 Speaker 4: That's right, that's just a rule. 158 00:09:39,800 --> 00:09:40,280 Speaker 2: That's just a. 159 00:09:40,320 --> 00:09:43,600 Speaker 5: Rule, right, and you're able to converse and yeah, I 160 00:09:43,640 --> 00:09:46,640 Speaker 5: need to book holiday, I need to move somebody between managers. 161 00:09:46,840 --> 00:09:49,640 Speaker 5: I need to figure out the policy on this. And 162 00:09:49,720 --> 00:09:53,400 Speaker 5: the AI basically navigates across the different systems to be 163 00:09:53,440 --> 00:09:56,440 Speaker 5: able to help get that information, to summarize it back, 164 00:09:56,920 --> 00:09:59,040 Speaker 5: to be able to carry out the transactions that I 165 00:09:59,080 --> 00:10:01,760 Speaker 5: need carried out, and it just just removes all of 166 00:10:01,760 --> 00:10:05,800 Speaker 5: that complexity and makes it easier to get things done. 167 00:10:06,520 --> 00:10:13,640 Speaker 3: When you are working with companies to implement generative AI. Now, 168 00:10:14,360 --> 00:10:17,720 Speaker 3: what do you find tends to be their primary focus? 169 00:10:18,400 --> 00:10:20,680 Speaker 4: I mean I speak with a lot of customers each week, 170 00:10:20,760 --> 00:10:25,000 Speaker 4: and for the last several months, most organizations have just 171 00:10:25,080 --> 00:10:29,280 Speaker 4: been reorienting themselves in terms of where are we in 172 00:10:29,280 --> 00:10:32,880 Speaker 4: this moment, what is this technology capable of? What are 173 00:10:32,920 --> 00:10:36,520 Speaker 4: the risks and governance and frameworks that I need to 174 00:10:36,880 --> 00:10:41,040 Speaker 4: establish in order to engage and talk to everyone. Talk 175 00:10:41,040 --> 00:10:44,040 Speaker 4: to my vendors, talk to my cloud providers, talk to 176 00:10:44,120 --> 00:10:48,960 Speaker 4: my consultants, talk to academics, and generally get your sea 177 00:10:49,000 --> 00:10:53,440 Speaker 4: legs under them. And the sort of the unstructured hand 178 00:10:53,480 --> 00:10:57,680 Speaker 4: on keyboards fiddling with technology seems to be moving towards 179 00:10:57,880 --> 00:11:00,760 Speaker 4: let's get some points on the board, turn this stuff 180 00:11:00,800 --> 00:11:03,320 Speaker 4: on and go. So that's what I've been seeing in 181 00:11:03,400 --> 00:11:07,440 Speaker 4: terms of, you know, the work within the salesforce ecosystem. Matt, 182 00:11:07,480 --> 00:11:11,199 Speaker 4: you've got a larger aperture as well. What are you seeing? 183 00:11:11,800 --> 00:11:14,400 Speaker 6: Yeah, so I definitely agree. 184 00:11:14,440 --> 00:11:17,959 Speaker 5: I think, you know, there's been lots of getting sea 185 00:11:18,040 --> 00:11:21,520 Speaker 5: legs experimentation, just trying to build knowledge, being able to 186 00:11:21,559 --> 00:11:27,040 Speaker 5: try and build almost you know, internal organizational point of 187 00:11:27,080 --> 00:11:29,959 Speaker 5: view and reference framework. I've seen lots of what I 188 00:11:30,000 --> 00:11:32,040 Speaker 5: would have referred to as random acts of AI. 189 00:11:34,000 --> 00:11:36,359 Speaker 6: In terms of in terms of experimentation. 190 00:11:36,440 --> 00:11:38,559 Speaker 5: But I think I think people now looking into twenty 191 00:11:38,600 --> 00:11:41,680 Speaker 5: twenty four and this is all about now adoption and scaling. 192 00:11:42,280 --> 00:11:46,160 Speaker 5: What's become really clear is organizations have started to realize 193 00:11:46,200 --> 00:11:48,280 Speaker 5: this is going to be a very multi model world 194 00:11:48,400 --> 00:11:49,880 Speaker 5: that they're going to live in. There is no one 195 00:11:49,920 --> 00:11:53,719 Speaker 5: AI that is the answer for their organization, and so 196 00:11:53,960 --> 00:11:57,360 Speaker 5: they're going to have lots of different generative AI models 197 00:11:58,040 --> 00:12:00,640 Speaker 5: and technologies that they're going to sit in the organization 198 00:12:00,920 --> 00:12:05,880 Speaker 5: servicing different use cases, different domain areas, different products and services, 199 00:12:07,160 --> 00:12:10,040 Speaker 5: and so therefore having to figure out how they're going 200 00:12:10,080 --> 00:12:13,160 Speaker 5: to navigate and manage this kind of open world that 201 00:12:13,160 --> 00:12:15,439 Speaker 5: they're going to be sitting in and the decisions that 202 00:12:15,480 --> 00:12:17,920 Speaker 5: they're going to have to make around that. I think 203 00:12:17,960 --> 00:12:21,160 Speaker 5: the second thing that I've seen that people are now 204 00:12:21,200 --> 00:12:23,640 Speaker 5: becoming very clear that this needs to be what I 205 00:12:23,679 --> 00:12:26,400 Speaker 5: would refer to as use case lead and outcome focused, 206 00:12:27,679 --> 00:12:31,080 Speaker 5: and so really needing to start with thinking about the 207 00:12:31,120 --> 00:12:34,120 Speaker 5: business outcome and the problem that you know we're trying 208 00:12:34,200 --> 00:12:38,199 Speaker 5: to solve, and therefore, how do I use generative AI 209 00:12:38,880 --> 00:12:42,280 Speaker 5: as part of the mechanism to solve that problem? And 210 00:12:42,320 --> 00:12:44,960 Speaker 5: I think you know what Susan and the Salesforce team 211 00:12:45,000 --> 00:12:46,840 Speaker 5: do is an amazing example of that. You know, they've 212 00:12:46,840 --> 00:12:51,240 Speaker 5: got this incredible platform and engine that allows companies to 213 00:12:51,280 --> 00:12:54,319 Speaker 5: transform their sales and service processes and to be able 214 00:12:54,320 --> 00:12:56,360 Speaker 5: to put data in the hands of users, to be 215 00:12:56,400 --> 00:12:59,440 Speaker 5: able to make better decisions, et cetera. And so now 216 00:12:59,480 --> 00:13:02,320 Speaker 5: by weaving generative AI into that platform, we're going to 217 00:13:02,320 --> 00:13:05,319 Speaker 5: be able to make those processes workflows even more efficient. Right, 218 00:13:05,360 --> 00:13:08,319 Speaker 5: So it's generative AI plus all of these other amazing 219 00:13:08,360 --> 00:13:10,160 Speaker 5: things that are there, but it will be led through 220 00:13:10,360 --> 00:13:13,679 Speaker 5: business outcome, and it will be led through use case 221 00:13:13,800 --> 00:13:16,760 Speaker 5: and the business problem or workflow that we're trying to improve. 222 00:13:17,520 --> 00:13:19,360 Speaker 5: And then I think the third thing is shifting from 223 00:13:19,360 --> 00:13:22,280 Speaker 5: this experimentation to scale. You know, I think everybody's really 224 00:13:22,320 --> 00:13:26,679 Speaker 5: early in this journey, but what's become clear is that 225 00:13:26,800 --> 00:13:30,840 Speaker 5: you know, everybody now need realizes and is starting to 226 00:13:30,920 --> 00:13:35,360 Speaker 5: lay down these these ground rules, the guardrails, the frameworks 227 00:13:35,400 --> 00:13:39,760 Speaker 5: to allow them to scale this across the organization. So, 228 00:13:40,600 --> 00:13:43,080 Speaker 5: you know, I think we're in for an exciting, exciting 229 00:13:43,120 --> 00:13:44,400 Speaker 5: time in twenty twenty four. 230 00:13:45,160 --> 00:13:47,880 Speaker 3: So now that we're getting to this moment, what are 231 00:13:47,880 --> 00:13:51,199 Speaker 3: the key things companies have to figure out about scaling 232 00:13:51,280 --> 00:13:51,920 Speaker 3: generative AI? 233 00:13:54,360 --> 00:13:57,600 Speaker 4: I would put that in kind of two categories and 234 00:13:57,679 --> 00:14:01,320 Speaker 4: following on what Matt was saying in terms of use, 235 00:14:01,360 --> 00:14:04,800 Speaker 4: case defined and outcome lead one hundred percent on that 236 00:14:04,880 --> 00:14:07,600 Speaker 4: in terms of starting with a hypothesis of value, while 237 00:14:07,600 --> 00:14:11,480 Speaker 4: at the same time people are getting closer to the 238 00:14:11,520 --> 00:14:14,280 Speaker 4: technology to know what their bounds are. But the biggest 239 00:14:14,360 --> 00:14:18,480 Speaker 4: you know, set of conversations is in the enterprise area 240 00:14:19,520 --> 00:14:22,800 Speaker 4: in terms of embarking and using with generative AI, how 241 00:14:22,800 --> 00:14:26,560 Speaker 4: to do it in ways that is safe for use 242 00:14:26,600 --> 00:14:30,560 Speaker 4: of data that is safe around not just the larger 243 00:14:30,640 --> 00:14:35,440 Speaker 4: topic of generative AI and hallucinations, which which are fun 244 00:14:35,480 --> 00:14:36,640 Speaker 4: to talk about in the media. 245 00:14:36,800 --> 00:14:39,560 Speaker 3: But it's a fun word, right. If it was called 246 00:14:39,600 --> 00:14:42,720 Speaker 3: something other than hallucinations, people wouldn't talk about it as much. 247 00:14:42,840 --> 00:14:46,360 Speaker 4: It was just mistakes, Yeah, that's right, just things that 248 00:14:46,400 --> 00:14:48,880 Speaker 4: aren't factually true. We've been doing a lot of work 249 00:14:48,880 --> 00:14:53,280 Speaker 4: at Salesforce around using you know, dynamic and structured grounding 250 00:14:53,320 --> 00:14:56,360 Speaker 4: the data so we can give very strong and non 251 00:14:56,480 --> 00:15:00,440 Speaker 4: naive prompt instructions to lllms to get return on that. So, 252 00:15:00,920 --> 00:15:04,680 Speaker 4: just to summarize, top of mind for organizations using you know, 253 00:15:04,960 --> 00:15:08,160 Speaker 4: large language models is using their data in ways that 254 00:15:08,240 --> 00:15:13,520 Speaker 4: are safe, trusted, not exposed, and reducing the opportunity for 255 00:15:13,560 --> 00:15:16,640 Speaker 4: hallucinations and maximizing relevant content. 256 00:15:17,200 --> 00:15:17,480 Speaker 5: Great. 257 00:15:17,520 --> 00:15:19,800 Speaker 3: So, so Matt Susan was talking about, you know, both 258 00:15:19,840 --> 00:15:23,760 Speaker 3: what organizations are concerned with as they scale generative AI 259 00:15:23,960 --> 00:15:27,680 Speaker 3: and how Salesforce is working to sort of address those concerns. 260 00:15:28,240 --> 00:15:31,200 Speaker 3: What are you seeing at IBM. 261 00:15:30,720 --> 00:15:34,360 Speaker 5: Here, So I think certainly from a scaling of generative 262 00:15:34,360 --> 00:15:39,560 Speaker 5: AI perspective, you know this topic of governance, you know, 263 00:15:39,600 --> 00:15:41,600 Speaker 5: and how organizations are going to have to govern all 264 00:15:41,600 --> 00:15:45,400 Speaker 5: of these models that sit withinside, how they manage kind 265 00:15:45,400 --> 00:15:49,760 Speaker 5: of bias fairness, model drift, you know, if you think 266 00:15:49,760 --> 00:15:53,080 Speaker 5: about the data that's gone into a model and the 267 00:15:53,120 --> 00:15:56,480 Speaker 5: output it gives to start with, not because the model changes, 268 00:15:56,480 --> 00:15:58,760 Speaker 5: but because the context of the world moves on. And 269 00:15:58,800 --> 00:16:01,080 Speaker 5: so being able to kind of manage this model drift 270 00:16:01,160 --> 00:16:03,160 Speaker 5: is going to be a really important thing. I think 271 00:16:03,240 --> 00:16:08,760 Speaker 5: data really matters, and so quality access security around data 272 00:16:08,800 --> 00:16:11,040 Speaker 5: within the enterprise is going to be critical to scaling 273 00:16:11,080 --> 00:16:13,560 Speaker 5: generative AI. And the other one I think that's going 274 00:16:13,560 --> 00:16:16,120 Speaker 5: to be really important, and I think many organizations haven't 275 00:16:16,120 --> 00:16:18,080 Speaker 5: even got there yet in their thinking is around the 276 00:16:18,320 --> 00:16:22,120 Speaker 5: ESG implications. So carbon you know, the use of this 277 00:16:22,200 --> 00:16:25,560 Speaker 5: technology does not come without a cost of carbon. 278 00:16:25,880 --> 00:16:28,640 Speaker 3: Carbon meaning it's very energy intensive. 279 00:16:28,520 --> 00:16:32,120 Speaker 5: Correct, Yeah, the training of the models and so thinking 280 00:16:32,120 --> 00:16:36,200 Speaker 5: about carbon disclosures and thinking about where I'm infusing it 281 00:16:36,240 --> 00:16:38,560 Speaker 5: into my business and how much I'm using it and 282 00:16:38,600 --> 00:16:41,560 Speaker 5: what the carbon cost of that is. As I think 283 00:16:41,600 --> 00:16:46,960 Speaker 5: about the you know, you know, my own organizational responsibilities 284 00:16:46,960 --> 00:16:49,000 Speaker 5: to reduce carbon I think, you know, there's all of 285 00:16:49,040 --> 00:16:51,240 Speaker 5: these things that I think are going to become important 286 00:16:51,280 --> 00:16:54,320 Speaker 5: factors as people are thinking about the scaling implications of 287 00:16:54,360 --> 00:16:55,160 Speaker 5: this technology. 288 00:16:56,080 --> 00:16:59,560 Speaker 2: AI is already making new experiences possible, but we must 289 00:16:59,600 --> 00:17:02,600 Speaker 2: be in mind in how we integrate this new technology 290 00:17:02,760 --> 00:17:07,119 Speaker 2: as we continue scaling generative AI. Matt touched on some 291 00:17:07,200 --> 00:17:11,520 Speaker 2: crucial aspects from an IBM perspective. Governance, bias, fairness, and 292 00:17:11,520 --> 00:17:16,040 Speaker 2: security are all key considerations when organizations aim to expand 293 00:17:16,080 --> 00:17:21,400 Speaker 2: their use of generative AI. The environmental aspect is especially important, 294 00:17:21,960 --> 00:17:25,040 Speaker 2: and it's refreshing to hear leading thinkers like Matt and 295 00:17:25,119 --> 00:17:30,240 Speaker 2: Susan highlight these issues. As this technology continues to evolve, 296 00:17:30,720 --> 00:17:35,640 Speaker 2: these factors are becoming increasingly important for organizations to address. 297 00:17:36,520 --> 00:17:40,679 Speaker 2: The Historic collaboration between IBM and Salesforce is helping to 298 00:17:40,800 --> 00:17:44,159 Speaker 2: remedy issues companies face when scaling AI. 299 00:17:45,320 --> 00:17:50,240 Speaker 3: So IBM and Salesforce recently announced a new collaborative project 300 00:17:50,400 --> 00:17:53,280 Speaker 3: around generative AI. Tell me more about that. 301 00:17:54,520 --> 00:17:59,240 Speaker 5: We've been partners for over two decades now IBM and Salesforce, 302 00:17:59,280 --> 00:18:02,359 Speaker 5: and so within our consulting business, you know, we work 303 00:18:02,400 --> 00:18:05,919 Speaker 5: with Salesforce technology to help our clients implement that technology 304 00:18:05,960 --> 00:18:09,640 Speaker 5: to transform their businesses. We've got a huge practice, over 305 00:18:09,800 --> 00:18:14,000 Speaker 5: twelve thousand people with certifications around Salesforce platforms, and so 306 00:18:14,760 --> 00:18:16,760 Speaker 5: you know, as Susan and her team and the broader 307 00:18:16,760 --> 00:18:20,120 Speaker 5: team in Salesforce are infusing more capability into the platform 308 00:18:20,160 --> 00:18:23,480 Speaker 5: around generative AI, then our mission is really simple. It's 309 00:18:23,560 --> 00:18:27,680 Speaker 5: to help clients who are using the Salesforce platform adopt 310 00:18:27,720 --> 00:18:29,200 Speaker 5: those capabilities to help. 311 00:18:29,040 --> 00:18:31,240 Speaker 6: Them get more benefit within their organization. 312 00:18:31,800 --> 00:18:34,960 Speaker 5: You know, we're also a significant user of Salesforce technology 313 00:18:35,000 --> 00:18:38,520 Speaker 5: within IBM. We're one of Salesforce's largest customers globally, and 314 00:18:38,600 --> 00:18:41,400 Speaker 5: so you know, as we continue to transform our own 315 00:18:41,480 --> 00:18:45,560 Speaker 5: sales and service processes within IBM, then you know, our 316 00:18:45,680 --> 00:18:48,760 Speaker 5: use of the generative AI capabilities that they're infusing into 317 00:18:48,800 --> 00:18:52,399 Speaker 5: sales cloud, service, cloud slack, et cetera will be something 318 00:18:52,400 --> 00:18:56,000 Speaker 5: that will become really important to us driving productivity within 319 00:18:56,080 --> 00:18:58,160 Speaker 5: the company. And then the other thing that I would 320 00:18:58,160 --> 00:19:00,160 Speaker 5: say is, you know, as I think about the work 321 00:19:00,160 --> 00:19:03,119 Speaker 5: that we do with clients, you know, as they're implementing 322 00:19:03,119 --> 00:19:05,639 Speaker 5: and on their generative AI journeys, you know, they're going 323 00:19:05,680 --> 00:19:09,600 Speaker 5: to utilize and leverage the salesforce capabilities within the platform 324 00:19:09,640 --> 00:19:13,240 Speaker 5: and their generative AI technologies. But then you start thinking 325 00:19:13,280 --> 00:19:16,840 Speaker 5: about processes and workflows that run beyond the walls of CRM, 326 00:19:16,920 --> 00:19:20,120 Speaker 5: right that run into supply chain and into the finance 327 00:19:20,200 --> 00:19:23,280 Speaker 5: area of the organization. And so there is work that 328 00:19:23,320 --> 00:19:26,080 Speaker 5: we're doing with clients where we're using ibms. What's the 329 00:19:26,160 --> 00:19:30,120 Speaker 5: next platform to be able to help get access to 330 00:19:30,119 --> 00:19:33,440 Speaker 5: to generate insights from data sources that sit in all 331 00:19:33,480 --> 00:19:35,680 Speaker 5: of these kind of back office areas of the enterprise, 332 00:19:36,240 --> 00:19:38,399 Speaker 5: and to be able to get that data across the 333 00:19:38,480 --> 00:19:43,040 Speaker 5: salesforce into these customer interaction points and into the employees 334 00:19:43,080 --> 00:19:47,280 Speaker 5: who are servicing those customers using salesforces AI and generative 335 00:19:47,280 --> 00:19:48,199 Speaker 5: AI technologies. 336 00:19:48,200 --> 00:19:49,160 Speaker 6: So there's a. 337 00:19:49,160 --> 00:19:52,080 Speaker 5: Kind of one plus one equals three kind of you know, 338 00:19:52,240 --> 00:19:55,399 Speaker 5: better together, you know, and being able to bring our 339 00:19:55,440 --> 00:19:59,520 Speaker 5: technologies together in service of these clients. Problems as you 340 00:19:59,560 --> 00:20:03,680 Speaker 5: think about these processes that run across their enterprise. So, yeah, 341 00:20:03,840 --> 00:20:06,560 Speaker 5: it's so huge hutunity and what we're doing together in 342 00:20:06,560 --> 00:20:08,000 Speaker 5: the market to help clients. 343 00:20:08,600 --> 00:20:10,879 Speaker 4: Yeah, and just building it on that. It is a 344 00:20:10,960 --> 00:20:15,840 Speaker 4: huge moment for organizations and for technology companies like Salesforce, 345 00:20:15,840 --> 00:20:18,639 Speaker 4: and we couldn't be happier to have partnerships like we 346 00:20:18,720 --> 00:20:23,439 Speaker 4: have with IBM. Like the range of thought leadership that 347 00:20:24,280 --> 00:20:27,679 Speaker 4: is appropriate at the moment is everything from what is 348 00:20:27,680 --> 00:20:30,720 Speaker 4: that hypothesis of value and what are those use cases? 349 00:20:30,760 --> 00:20:33,240 Speaker 4: And what is the order of operation in terms of 350 00:20:33,280 --> 00:20:37,120 Speaker 4: approaching it just in terms of focus, but then things 351 00:20:37,200 --> 00:20:41,480 Speaker 4: that would help organizations assess their AI readiness and then 352 00:20:41,520 --> 00:20:44,719 Speaker 4: their approach like you know, we talked earlier about frameworks 353 00:20:44,760 --> 00:20:48,040 Speaker 4: and guardrails. You know, what are use cases that we're 354 00:20:48,080 --> 00:20:50,920 Speaker 4: comfortable with given the state of the technology that face 355 00:20:50,960 --> 00:20:55,160 Speaker 4: employees or face customers. So creating these much larger roadmaps 356 00:20:55,200 --> 00:20:58,000 Speaker 4: in terms of how to approach this over a series 357 00:20:58,080 --> 00:21:03,400 Speaker 4: of initiatives, it can fundamentally change the way we engage 358 00:21:03,400 --> 00:21:07,800 Speaker 4: with technology and what that means for the you know, 359 00:21:07,880 --> 00:21:13,000 Speaker 4: training and change management and use cases that fundamentally shift 360 00:21:13,160 --> 00:21:17,080 Speaker 4: how you engage with systems like salesforces. There's just a 361 00:21:17,080 --> 00:21:19,040 Speaker 4: massive opportunity for us together. 362 00:21:19,840 --> 00:21:23,680 Speaker 3: So you're talking in sort of general terms, I'm interested in, 363 00:21:23,840 --> 00:21:28,240 Speaker 3: you know, thinking in particular about the way generitive AI 364 00:21:28,359 --> 00:21:32,280 Speaker 3: can essentially lead to better business outcomes, right Like, what 365 00:21:32,320 --> 00:21:34,840 Speaker 3: does that look like? How do you measure it? You know, 366 00:21:34,960 --> 00:21:37,480 Speaker 3: there's a certain bottom line question there, right like, how 367 00:21:37,480 --> 00:21:40,000 Speaker 3: does AI make businesses work better? And in what ways? 368 00:21:40,640 --> 00:21:44,920 Speaker 4: You know, as consumers of products and services, we all 369 00:21:44,960 --> 00:21:47,560 Speaker 4: love and respect great service, you know, in terms of 370 00:21:47,560 --> 00:21:50,920 Speaker 4: getting timely, quick answers, resolving issues quickly, all those those 371 00:21:50,960 --> 00:21:55,920 Speaker 4: types of things. And from the perspective of using generative 372 00:21:55,960 --> 00:22:00,160 Speaker 4: and predictive capabilities for agents who are interacting with customers, 373 00:22:00,520 --> 00:22:03,760 Speaker 4: there is just a whole ton of opportunity to take 374 00:22:03,760 --> 00:22:06,280 Speaker 4: friction out of the process in terms of finding answers, 375 00:22:06,320 --> 00:22:10,000 Speaker 4: resolving issues, in terms of using these generative capabilities that 376 00:22:10,080 --> 00:22:13,200 Speaker 4: will bring you know, answers and content to the fingertips 377 00:22:13,240 --> 00:22:18,480 Speaker 4: more easily to the human agents that are working with customers. Now, 378 00:22:18,600 --> 00:22:22,200 Speaker 4: taking that to the next step for organizations when they're 379 00:22:22,240 --> 00:22:26,040 Speaker 4: ready to move into more customer facing automation, that's yet 380 00:22:26,080 --> 00:22:28,919 Speaker 4: another channel. As a consumer, we'll all enjoy with the 381 00:22:28,920 --> 00:22:31,120 Speaker 4: brands and the products and the services that we want 382 00:22:31,160 --> 00:22:34,679 Speaker 4: in terms of fast answers and resolutions to customers. And 383 00:22:34,720 --> 00:22:39,480 Speaker 4: as we all know, great customer experience yields return business. 384 00:22:40,000 --> 00:22:43,240 Speaker 4: Now on the sales side, you know, maybe a different example, 385 00:22:43,800 --> 00:22:46,760 Speaker 4: and these are areas where I think the capability of 386 00:22:46,920 --> 00:22:50,439 Speaker 4: predictive and generative go very well together in terms of 387 00:22:50,480 --> 00:22:54,280 Speaker 4: focusing on business outcomes. And a classic example would be, 388 00:22:54,800 --> 00:22:59,159 Speaker 4: you know, predictions that help us understand customer health. You know, 389 00:22:59,280 --> 00:23:03,200 Speaker 4: is this customer engaged, is this customer at risk? Predictions 390 00:23:03,240 --> 00:23:07,359 Speaker 4: that help us understand next best product or next best conversation. 391 00:23:07,880 --> 00:23:13,520 Speaker 4: These all help focus sales team's time on a customer 392 00:23:13,600 --> 00:23:17,200 Speaker 4: or a territory, and so that deep focus puts all 393 00:23:17,200 --> 00:23:19,520 Speaker 4: the wood behind an arrow, so to speak, in terms 394 00:23:19,560 --> 00:23:24,000 Speaker 4: of where we should be engaging. And those types of 395 00:23:24,520 --> 00:23:29,199 Speaker 4: driven sales organizations that have these capabilities just lead to 396 00:23:29,240 --> 00:23:34,080 Speaker 4: better performance and outcomes and customer experience too. Now, let's 397 00:23:34,119 --> 00:23:38,600 Speaker 4: also layer in generitive capabilities where we're using the generative 398 00:23:38,640 --> 00:23:42,680 Speaker 4: capabilities to assist and augment a sales team, where we're 399 00:23:42,720 --> 00:23:46,800 Speaker 4: using the power de generitive for everything like generating personalized 400 00:23:46,880 --> 00:23:51,720 Speaker 4: and relevant customer interaction content, for example, leveraging our customer 401 00:23:51,800 --> 00:23:56,720 Speaker 4: data like engagement history, product purchases, service history to create 402 00:23:56,760 --> 00:23:59,760 Speaker 4: an email or a campaign. And this scale a lout 403 00:23:59,760 --> 00:24:03,119 Speaker 4: of has just never been possible before. And you know, 404 00:24:03,160 --> 00:24:05,640 Speaker 4: maybe even taking this one step further re genitive, where 405 00:24:05,720 --> 00:24:08,200 Speaker 4: we take all the administrative friction out of the day 406 00:24:08,280 --> 00:24:11,640 Speaker 4: job and doing things for sales teams like summarizing their 407 00:24:11,680 --> 00:24:15,040 Speaker 4: calls or creating a meeting plan for them, and you know, 408 00:24:15,200 --> 00:24:18,880 Speaker 4: very broadly speaking, using generative AI to change the interaction 409 00:24:19,000 --> 00:24:23,760 Speaker 4: mode with systems like Salesforce from clicks and training where 410 00:24:23,800 --> 00:24:26,840 Speaker 4: people have to focus on the process to more conversational 411 00:24:27,000 --> 00:24:31,560 Speaker 4: user experiences which are much more engaging and easier to use. 412 00:24:32,040 --> 00:24:35,960 Speaker 4: So all of this together is just incredible and transformational 413 00:24:36,359 --> 00:24:39,399 Speaker 4: and makes all businesses and people work better. 414 00:24:40,000 --> 00:24:42,480 Speaker 3: So I just want to spend one more moment on 415 00:24:42,560 --> 00:24:48,000 Speaker 3: the partnership between IBM and Salesforce and genitive AI. And 416 00:24:48,040 --> 00:24:52,200 Speaker 3: there's this phrase that's interesting to me. It's ecosystem partnership 417 00:24:52,480 --> 00:24:55,280 Speaker 3: that I think is relevant here. So what is an 418 00:24:55,320 --> 00:24:58,399 Speaker 3: ecosystem partnership and why is it you know, helpful in 419 00:25:00,280 --> 00:25:02,000 Speaker 3: scalable AI solutions. 420 00:25:02,880 --> 00:25:06,959 Speaker 5: This idea of being open I think is probably one 421 00:25:07,000 --> 00:25:11,200 Speaker 5: of the most important premises for US as technology companies, 422 00:25:11,320 --> 00:25:15,000 Speaker 5: for US as consultancies and system integrators, and for our 423 00:25:15,040 --> 00:25:18,840 Speaker 5: clients to think about that the sources of value that 424 00:25:18,920 --> 00:25:22,920 Speaker 5: can be created through taking an open approach is hugely important. 425 00:25:23,000 --> 00:25:26,840 Speaker 5: So if I think about for US, ecosystem means making 426 00:25:26,880 --> 00:25:31,240 Speaker 5: sure that we have all of the different partnerships that 427 00:25:31,280 --> 00:25:35,639 Speaker 5: we need with technology providers, with service providers that we 428 00:25:35,760 --> 00:25:41,080 Speaker 5: can bring to our clients the right set of capabilities 429 00:25:41,119 --> 00:25:43,400 Speaker 5: to solve the problem that they've got and not thinking 430 00:25:43,560 --> 00:25:46,960 Speaker 5: that just you know, what we have in house, or 431 00:25:46,960 --> 00:25:49,040 Speaker 5: what we have with just one other partner that we 432 00:25:49,119 --> 00:25:51,359 Speaker 5: work with, you know, is the right thing. And so 433 00:25:51,600 --> 00:25:54,280 Speaker 5: you know, I think every problem that our clients have 434 00:25:54,480 --> 00:25:58,000 Speaker 5: is solved through a range of technologies that come together 435 00:25:58,640 --> 00:26:01,320 Speaker 5: in service of creating that business outcome. 436 00:26:01,800 --> 00:26:07,159 Speaker 3: I want to touch briefly on ethics and governance. Something 437 00:26:07,280 --> 00:26:12,240 Speaker 3: like eighty percent of CEOs see explainability, ethics, bias, trust 438 00:26:12,480 --> 00:26:16,879 Speaker 3: as major concerns on the road to AI adoption, and 439 00:26:16,920 --> 00:26:21,840 Speaker 3: so I'm curious how business leaders navigate these things, and 440 00:26:21,880 --> 00:26:26,440 Speaker 3: in particular, how Salesforce and IBM are building these concerns 441 00:26:26,480 --> 00:26:29,000 Speaker 3: into how they work with customers. 442 00:26:29,520 --> 00:26:34,520 Speaker 4: You know, we've been incorporating predictive machine learning into our 443 00:26:34,600 --> 00:26:38,280 Speaker 4: products since mid last decade, and at that time we 444 00:26:38,400 --> 00:26:42,240 Speaker 4: started with all of our ethics and governance work at 445 00:26:42,280 --> 00:26:45,520 Speaker 4: that time in terms of frameworks for engaging with AI 446 00:26:46,000 --> 00:26:49,399 Speaker 4: in ethical and safe ways and have a lot of 447 00:26:49,400 --> 00:26:52,920 Speaker 4: guidance for customers in terms of those programs. The machine 448 00:26:52,960 --> 00:26:56,280 Speaker 4: learning focus that we've had at Salesforce has always been 449 00:26:56,400 --> 00:27:00,679 Speaker 4: deeply focused on explainability. So if we're making you know, 450 00:27:00,760 --> 00:27:05,199 Speaker 4: predictive recommendations to explain how we got to that, you know, 451 00:27:05,240 --> 00:27:08,680 Speaker 4: whether that's something that a user sees, is they're engaging 452 00:27:08,680 --> 00:27:11,280 Speaker 4: with it so they have full trust in terms of 453 00:27:11,400 --> 00:27:15,120 Speaker 4: interacting with it, but also for the practitioners who are 454 00:27:15,119 --> 00:27:18,560 Speaker 4: building it. So we have this like long standing vibe 455 00:27:18,560 --> 00:27:22,600 Speaker 4: and capability with our predictive side of the house and 456 00:27:22,720 --> 00:27:24,960 Speaker 4: on the generative side of the house. You know, the 457 00:27:25,000 --> 00:27:28,520 Speaker 4: state of the marketplace right now is llms for most 458 00:27:28,520 --> 00:27:32,960 Speaker 4: people are are largely black boxes in terms of not 459 00:27:33,119 --> 00:27:35,399 Speaker 4: fully interpretable in terms of how they come up with 460 00:27:35,440 --> 00:27:38,600 Speaker 4: their content. Now that said, there is a lot that 461 00:27:38,680 --> 00:27:42,600 Speaker 4: you can do in terms of audit, in terms of 462 00:27:42,760 --> 00:27:46,280 Speaker 4: you know, transparency in terms of what are the prompts 463 00:27:46,320 --> 00:27:49,920 Speaker 4: that are being submitted to these llms, what do these 464 00:27:50,040 --> 00:27:53,840 Speaker 4: llms provide back in terms of return? And then what 465 00:27:53,880 --> 00:27:55,960 Speaker 4: did the human do to change it, use it, or 466 00:27:56,280 --> 00:27:59,160 Speaker 4: adjust it. So we've been updating all of our ethics 467 00:27:59,240 --> 00:28:02,280 Speaker 4: and government it's frameworks now, I guess I would call 468 00:28:02,320 --> 00:28:05,040 Speaker 4: it with safety components as well in terms of how 469 00:28:05,040 --> 00:28:08,680 Speaker 4: to work with data in safe ways and with these 470 00:28:08,720 --> 00:28:10,880 Speaker 4: trened parents governance models. Yeah. 471 00:28:10,960 --> 00:28:12,959 Speaker 5: So, I mean this is an area that IBM has 472 00:28:13,000 --> 00:28:15,359 Speaker 5: been kind of working on for many years. And so 473 00:28:15,680 --> 00:28:18,480 Speaker 5: you know, our AI Ethics Board that we have internally 474 00:28:19,200 --> 00:28:23,000 Speaker 5: kind of governs and provides frameworks and guidance for everything 475 00:28:23,000 --> 00:28:25,159 Speaker 5: that we do in the company. There's a lot of 476 00:28:25,160 --> 00:28:28,480 Speaker 5: work that we do to help our clients and organizations 477 00:28:28,600 --> 00:28:32,199 Speaker 5: establish their strategies for AI governance as well as their 478 00:28:32,240 --> 00:28:37,840 Speaker 5: own internal policies, models, approaches, ethics boards, et cetera. And so, 479 00:28:38,240 --> 00:28:41,280 Speaker 5: you know, helping them put in place these ground rules 480 00:28:41,320 --> 00:28:47,560 Speaker 5: and guardrails, organizational process changes, et cetera. I think is 481 00:28:47,600 --> 00:28:50,080 Speaker 5: a really important part of this scaling discussion that we 482 00:28:50,080 --> 00:28:52,400 Speaker 5: were having earlier, as people are going to be kind 483 00:28:52,400 --> 00:28:55,880 Speaker 5: of rolling out more of this technology internally, and then 484 00:28:55,920 --> 00:28:59,680 Speaker 5: I think there's a lot that organizations are going to 485 00:28:59,680 --> 00:29:02,080 Speaker 5: have to do to think about, especially in the generative world, 486 00:29:02,680 --> 00:29:05,320 Speaker 5: around all of the different types of models that they're using, 487 00:29:05,600 --> 00:29:09,120 Speaker 5: models that they're training and tuning and building, and how 488 00:29:09,160 --> 00:29:13,320 Speaker 5: they manage all of those for explainability and bias drift 489 00:29:13,400 --> 00:29:18,200 Speaker 5: and actually regulatory requirements, Like if you think about what's 490 00:29:18,280 --> 00:29:22,400 Speaker 5: happening around the world, there's different countries, the EUAI Act, 491 00:29:22,480 --> 00:29:25,400 Speaker 5: you know, there's lots of different regulatory requirements that are 492 00:29:25,400 --> 00:29:28,080 Speaker 5: going to be coming in and so for multinational companies 493 00:29:28,840 --> 00:29:33,480 Speaker 5: operating across multiple countries, how they're going to have to 494 00:29:33,560 --> 00:29:37,040 Speaker 5: make sure that they're complying with all of not only 495 00:29:37,080 --> 00:29:41,520 Speaker 5: their own internal policies, but the requirements of the country 496 00:29:42,240 --> 00:29:47,480 Speaker 5: as well as potentially industry regulatory requirements as well. 497 00:29:47,520 --> 00:29:48,680 Speaker 6: And so there's a lot. 498 00:29:48,520 --> 00:29:50,800 Speaker 5: That we are doing and going to be doing in 499 00:29:51,360 --> 00:29:55,240 Speaker 5: helping them manage complexity. But IBM has a very firm 500 00:29:55,320 --> 00:29:57,920 Speaker 5: view that we believe that this is all about regulating 501 00:29:57,960 --> 00:30:02,920 Speaker 5: AI risk, not ail rhythms, and so focusing on precision regulation, 502 00:30:03,280 --> 00:30:07,720 Speaker 5: so you know, use the bodies and regulatory bodies that 503 00:30:07,760 --> 00:30:11,880 Speaker 5: are out there to provide the control as opposed to 504 00:30:11,880 --> 00:30:13,360 Speaker 5: trying to regulate the technology. 505 00:30:14,160 --> 00:30:18,120 Speaker 3: So genitive AI is changing kind of absurdly quickly. Right, 506 00:30:18,160 --> 00:30:19,600 Speaker 3: a year and a half ago, we wouldn't have been 507 00:30:19,600 --> 00:30:23,160 Speaker 3: having this conversation. We're here today. Everything's happening now. I'm 508 00:30:23,160 --> 00:30:26,160 Speaker 3: curious what you both think about about the near term 509 00:30:26,240 --> 00:30:29,160 Speaker 3: future of genitive A. Right, if we came back in 510 00:30:29,320 --> 00:30:31,000 Speaker 3: a year, or let's say two years from now. If 511 00:30:31,000 --> 00:30:33,640 Speaker 3: we came back two years from now to talk about 512 00:30:33,640 --> 00:30:35,840 Speaker 3: the work you're doing in genitive AI, what would we 513 00:30:35,920 --> 00:30:36,520 Speaker 3: be talking about. 514 00:30:38,200 --> 00:30:42,240 Speaker 4: I use this example sometimes I have three kids, and 515 00:30:42,960 --> 00:30:46,760 Speaker 4: I don't think any of them have ever gone into 516 00:30:46,800 --> 00:30:49,680 Speaker 4: a bank to deposit a check. Right, They pull out 517 00:30:49,720 --> 00:30:52,640 Speaker 4: their mobile phone and they scan the check with the 518 00:30:52,680 --> 00:30:53,800 Speaker 4: camera and they're done. 519 00:30:53,920 --> 00:30:56,200 Speaker 3: I'm surprised that they even know what a check is. 520 00:30:56,040 --> 00:30:59,320 Speaker 4: For the record, but yeah, right, well, yeah, sometimes their 521 00:30:59,360 --> 00:31:03,800 Speaker 4: parents give them one, like they get direct deposit. But anyway, 522 00:31:03,880 --> 00:31:07,080 Speaker 4: like this experience of like, what do you mean I 523 00:31:07,200 --> 00:31:09,320 Speaker 4: go into a branch in cash a check. I just 524 00:31:09,400 --> 00:31:11,720 Speaker 4: do this with my mobile phone. And I think a 525 00:31:11,760 --> 00:31:14,040 Speaker 4: little bit of it that way, in terms of the 526 00:31:14,080 --> 00:31:17,520 Speaker 4: systems that we use at work. I can imagine explaining 527 00:31:17,560 --> 00:31:21,200 Speaker 4: to my kids like, oh yeah, at Salesforce. You know, 528 00:31:21,320 --> 00:31:23,640 Speaker 4: back when someone had their first day on the job, 529 00:31:24,000 --> 00:31:27,080 Speaker 4: you know, as a service agent or as a salesperson, 530 00:31:27,520 --> 00:31:29,920 Speaker 4: they would have tabs on the screen and they would 531 00:31:29,960 --> 00:31:33,360 Speaker 4: be trained where to click, and they'd have documented processes 532 00:31:33,520 --> 00:31:36,880 Speaker 4: in manuals and that showed them where to get from 533 00:31:36,920 --> 00:31:40,360 Speaker 4: point A to point B. And as the clock turns forward, 534 00:31:40,880 --> 00:31:44,960 Speaker 4: they're just interacting with the natural language prompt. But it 535 00:31:45,080 --> 00:31:49,160 Speaker 4: just kind of fundamentally changes the way we'll be able 536 00:31:49,200 --> 00:31:51,240 Speaker 4: to interact with our systems a record at work. 537 00:31:51,760 --> 00:31:55,080 Speaker 3: It'll be just much more conversational. Instead of clicking through something, 538 00:31:55,240 --> 00:31:57,600 Speaker 3: you'll just basically have a conversation. 539 00:31:57,400 --> 00:31:58,640 Speaker 4: Much more conversational. 540 00:31:58,800 --> 00:32:01,200 Speaker 5: Yeah, this is the biggest paradigm shift in how we 541 00:32:01,240 --> 00:32:03,840 Speaker 5: interact with technology, I think since the invention of the 542 00:32:03,880 --> 00:32:07,520 Speaker 5: graphical user interface, and it's going to enable us to 543 00:32:07,680 --> 00:32:11,960 Speaker 5: almost put aside all of that complexity within organizations around 544 00:32:12,080 --> 00:32:15,920 Speaker 5: system silos, process silos, flows, because you're just going to 545 00:32:16,000 --> 00:32:20,120 Speaker 5: layer this just simple natural language interface over all of 546 00:32:20,160 --> 00:32:21,040 Speaker 5: that complexity. 547 00:32:22,000 --> 00:32:23,000 Speaker 6: Yeah, it's going to. 548 00:32:22,960 --> 00:32:26,960 Speaker 5: Amplify, i think the potential of every person on every 549 00:32:26,960 --> 00:32:29,400 Speaker 5: team in a way that we've never been able to 550 00:32:29,440 --> 00:32:32,160 Speaker 5: see before. And the other thing that I think as 551 00:32:32,240 --> 00:32:34,560 Speaker 5: you project forward in a couple of years, and Susan 552 00:32:34,640 --> 00:32:36,200 Speaker 5: just picking up on the point that you talked about 553 00:32:36,240 --> 00:32:37,800 Speaker 5: about blanking, you know. 554 00:32:37,840 --> 00:32:39,840 Speaker 6: I think there's a wonderful little example. 555 00:32:40,080 --> 00:32:41,800 Speaker 5: Look, if you think back to the seventies and the 556 00:32:41,840 --> 00:32:45,040 Speaker 5: eighties when the ATM kind of cash machines were rolling out, 557 00:32:45,720 --> 00:32:49,680 Speaker 5: and at that time, it wasn't really a reaction that 558 00:32:49,800 --> 00:32:52,520 Speaker 5: was one of awe or appreciation for convenience, but people 559 00:32:52,560 --> 00:32:55,880 Speaker 5: were concerned that we were automating away the bank teller jobs. 560 00:32:56,640 --> 00:32:58,280 Speaker 6: Right. But now, when. 561 00:32:58,160 --> 00:33:00,880 Speaker 5: You think about it, what actually had and was this 562 00:33:01,120 --> 00:33:05,360 Speaker 5: technology allowed the banks to scale their branch networks, more 563 00:33:05,400 --> 00:33:09,240 Speaker 5: branches never before, more bank tellers than ever before. Bank 564 00:33:09,320 --> 00:33:12,680 Speaker 5: teller employment and salaries increased, even though we automated them 565 00:33:12,680 --> 00:33:15,480 Speaker 5: out of work, because when they weren't having to spend 566 00:33:15,480 --> 00:33:18,400 Speaker 5: their time counting cash out for people, they were able 567 00:33:18,400 --> 00:33:20,680 Speaker 5: to do more valuable things, right, and new types of 568 00:33:20,720 --> 00:33:24,120 Speaker 5: financial products and services and mortgages and so like. If 569 00:33:24,160 --> 00:33:26,600 Speaker 5: I think back to that in the seventies and eighties 570 00:33:26,640 --> 00:33:28,840 Speaker 5: and then I project to where we are today, we're 571 00:33:28,840 --> 00:33:32,880 Speaker 5: just going to unleash this creativity and potential for employees 572 00:33:32,920 --> 00:33:35,680 Speaker 5: and enterprises by freeing up the time that they're spending 573 00:33:35,760 --> 00:33:38,000 Speaker 5: on things that you know, they can do far more 574 00:33:38,080 --> 00:33:40,200 Speaker 5: value added tasks. And so I think we're going to 575 00:33:40,200 --> 00:33:43,960 Speaker 5: be amazed I think around what happens and what companies 576 00:33:44,000 --> 00:33:45,440 Speaker 5: and people are going to be able to do as 577 00:33:45,480 --> 00:33:47,720 Speaker 5: we give them the time and space to be able 578 00:33:47,760 --> 00:33:49,040 Speaker 5: to do that great. 579 00:33:49,680 --> 00:33:52,840 Speaker 3: So, just to close, I want to talk about how 580 00:33:52,880 --> 00:33:56,400 Speaker 3: both of you use creativity in your own work. Just 581 00:33:56,440 --> 00:33:58,600 Speaker 3: to start with you, Matt, I know that you love 582 00:33:58,680 --> 00:34:04,560 Speaker 3: to combine create and technology through design. Do you use 583 00:34:04,640 --> 00:34:07,000 Speaker 3: generative AI in your own creative process? 584 00:34:07,400 --> 00:34:07,760 Speaker 6: Yeah? 585 00:34:07,840 --> 00:34:12,360 Speaker 5: So I'm a firm believer that this combination of experience 586 00:34:12,400 --> 00:34:14,799 Speaker 5: in AI is going to be the thing that makes 587 00:34:14,800 --> 00:34:18,000 Speaker 5: a difference. Like these large language models, and this technology 588 00:34:18,000 --> 00:34:20,680 Speaker 5: has been around actually for a number of years, and 589 00:34:20,719 --> 00:34:24,080 Speaker 5: it's only at the point late twenty twenty two where 590 00:34:24,360 --> 00:34:27,520 Speaker 5: open AI wrapped a digital experience around this and put 591 00:34:27,520 --> 00:34:30,319 Speaker 5: it in the hands of people that suddenly the transformative 592 00:34:30,360 --> 00:34:33,399 Speaker 5: power of this technology was realized. And so I think 593 00:34:33,440 --> 00:34:36,920 Speaker 5: the way that we surface these capabilities and put them 594 00:34:36,960 --> 00:34:40,160 Speaker 5: in the hands of people to be able to adopt 595 00:34:40,200 --> 00:34:42,960 Speaker 5: it in a really frictionless way is the thing that's 596 00:34:43,000 --> 00:34:45,520 Speaker 5: going to be hugely important to the adoption and. 597 00:34:45,560 --> 00:34:46,080 Speaker 6: Scaling of this. 598 00:34:46,200 --> 00:34:49,080 Speaker 5: So I think the most important thing for companies to 599 00:34:49,080 --> 00:34:53,120 Speaker 5: do is to make people, not technology central to their strategy. 600 00:34:53,800 --> 00:34:56,680 Speaker 3: Just to go more broadly into your works as a 601 00:34:56,840 --> 00:34:59,920 Speaker 3: I mean, I know that you have launched sales for 602 00:35:00,080 --> 00:35:03,000 Speaker 3: versus AI products into the market, and that you know 603 00:35:03,040 --> 00:35:05,960 Speaker 3: a lot of those have been built obviously given Salesforce 604 00:35:06,000 --> 00:35:10,840 Speaker 3: business around helping people build stronger customer relationships, right, and 605 00:35:10,920 --> 00:35:14,040 Speaker 3: so I'm curious what creativity did you bring to that work. 606 00:35:14,960 --> 00:35:17,480 Speaker 4: Some of the products that I've worked with Salesforce, they're 607 00:35:17,640 --> 00:35:22,080 Speaker 4: they're deeply visually focused. And my personal perspective is is 608 00:35:22,120 --> 00:35:26,600 Speaker 4: that the world can be really noisy. We're just inundated 609 00:35:27,239 --> 00:35:30,279 Speaker 4: with all sorts of demands on our time through so 610 00:35:30,400 --> 00:35:33,640 Speaker 4: many channels, right, Like the phone is firing off, you're 611 00:35:33,640 --> 00:35:37,719 Speaker 4: getting instant messages, you're getting slack messages, you're getting you know, DMS, 612 00:35:38,000 --> 00:35:41,440 Speaker 4: you're getting emails, your phone is ringing. There's processes that 613 00:35:41,480 --> 00:35:44,759 Speaker 4: are bearing down on you. And if we can use 614 00:35:44,840 --> 00:35:49,680 Speaker 4: really good design to filter out and essentially weed the garden, 615 00:35:50,200 --> 00:35:52,960 Speaker 4: because you know, we have this this phrase at Salesforces everything, 616 00:35:53,000 --> 00:35:57,320 Speaker 4: if everything's important, nothing's important. So using really good design 617 00:35:57,960 --> 00:36:02,680 Speaker 4: to create the user experience in salesforce, that just brings 618 00:36:02,880 --> 00:36:05,920 Speaker 4: stuff to life in the most powerful way. So I 619 00:36:05,960 --> 00:36:08,319 Speaker 4: always think of it from that perspective, like, if I'm 620 00:36:08,320 --> 00:36:11,759 Speaker 4: going to put this on a screen and salesforce, what 621 00:36:11,920 --> 00:36:14,799 Speaker 4: did I not put on? Is this the most important thing? 622 00:36:15,239 --> 00:36:17,400 Speaker 4: And is this the thing that's going to align everyone 623 00:36:17,440 --> 00:36:20,640 Speaker 4: to the larger initiative of the firm. So it's that 624 00:36:20,760 --> 00:36:24,520 Speaker 4: kind of design thinking that I use probably every moment 625 00:36:24,520 --> 00:36:27,760 Speaker 4: of the day, whether I'm building a demo or talking 626 00:36:27,760 --> 00:36:29,920 Speaker 4: to an executive as a company in terms of as 627 00:36:29,960 --> 00:36:31,920 Speaker 4: I see a vision for how they might deploy our 628 00:36:31,960 --> 00:36:34,600 Speaker 4: products to actual product development. 629 00:36:35,840 --> 00:36:38,040 Speaker 3: Just to kind of bring together these two themes we've 630 00:36:38,080 --> 00:36:40,040 Speaker 3: been talking about, on the one hand, the sort of 631 00:36:40,360 --> 00:36:44,439 Speaker 3: ecosystem partnerships and on the other hand, creativity. I mean, 632 00:36:44,719 --> 00:36:48,319 Speaker 3: can you talk a little bit about how working with 633 00:36:49,080 --> 00:36:52,480 Speaker 3: working with partners can foster a different kind of creativity. 634 00:36:53,160 --> 00:36:56,120 Speaker 4: More perspectives are always better than few perspectives. 635 00:36:56,760 --> 00:36:57,840 Speaker 6: I completely agree. 636 00:36:57,920 --> 00:37:01,879 Speaker 5: I think the mole minds, the more perspective, the more experiences. 637 00:37:03,200 --> 00:37:05,560 Speaker 5: You know, if I think about some of the best sessions, 638 00:37:05,760 --> 00:37:09,439 Speaker 5: best workshops, best work we do with clients. It's when 639 00:37:09,480 --> 00:37:13,680 Speaker 5: you've got people not just from one industry, but from 640 00:37:13,680 --> 00:37:17,680 Speaker 5: many industries, because actually the adjacencies and the things that 641 00:37:17,680 --> 00:37:20,560 Speaker 5: are happening in these other spaces trigger new thoughts and 642 00:37:20,600 --> 00:37:24,000 Speaker 5: new ideas. And so, you know, I think the richness 643 00:37:24,000 --> 00:37:27,960 Speaker 5: that we get when we partner with Salesforce together around 644 00:37:27,960 --> 00:37:32,239 Speaker 5: helping clients transform their front office, their sales service marketing processes. 645 00:37:32,600 --> 00:37:33,960 Speaker 6: We all bring these unique. 646 00:37:33,719 --> 00:37:36,759 Speaker 5: Experiences, and I think that just opens the aperture to 647 00:37:36,840 --> 00:37:40,680 Speaker 5: better outcomes and better perspectives for our clients. 648 00:37:41,640 --> 00:37:43,839 Speaker 4: Well, you know, you've been asking these questions about like 649 00:37:44,080 --> 00:37:47,239 Speaker 4: the use of tech and AI and creativity are sort 650 00:37:47,280 --> 00:37:49,360 Speaker 4: of in the same sentence. And one of the things 651 00:37:49,360 --> 00:37:53,160 Speaker 4: that I also think of is in terms of remaining 652 00:37:53,200 --> 00:37:57,719 Speaker 4: deeply creative is the actual process of unplugging from all 653 00:37:57,719 --> 00:38:02,520 Speaker 4: that stuff. So taking a trail run with no earphones 654 00:38:02,719 --> 00:38:06,680 Speaker 4: in your head, for me, is always a really good 655 00:38:06,680 --> 00:38:10,239 Speaker 4: way of unleashing and unbridening a lot of you know, 656 00:38:10,400 --> 00:38:14,960 Speaker 4: creative spirit. Just that downtime and the unstructured time where 657 00:38:14,960 --> 00:38:17,799 Speaker 4: your brain can just run free, actually not assisted by 658 00:38:17,800 --> 00:38:20,720 Speaker 4: any kind of device in my head or in my face. 659 00:38:20,880 --> 00:38:24,960 Speaker 3: So I think with that praise of unplugged time. We 660 00:38:25,000 --> 00:38:27,719 Speaker 3: should say goodbye and let's unplug it. It's lovely to 661 00:38:27,719 --> 00:38:29,920 Speaker 3: talk with you guys. It was really interesting to learn 662 00:38:29,960 --> 00:38:32,120 Speaker 3: about your work and the relationship between the company. So 663 00:38:32,239 --> 00:38:33,120 Speaker 3: thank you for your time. 664 00:38:33,960 --> 00:38:35,200 Speaker 6: Thank you, Jacob, thank you. 665 00:38:36,760 --> 00:38:39,080 Speaker 2: A huge thanks is due to Jacob, Matt and Susan 666 00:38:39,200 --> 00:38:44,399 Speaker 2: for illuminating the possibilities of generative AI. This technology has 667 00:38:44,440 --> 00:38:48,279 Speaker 2: great promise for creating new experiences in the future, but 668 00:38:48,480 --> 00:38:54,040 Speaker 2: requires the scaling capabilities made possible by partnerships like IBM 669 00:38:54,600 --> 00:38:58,839 Speaker 2: and Salesforce. As our conversation with Susan and Matt illustrated, 670 00:38:59,200 --> 00:39:03,440 Speaker 2: we're at an exiting phase of adoption. Most companies have 671 00:39:03,560 --> 00:39:08,200 Speaker 2: moved beyond experimentation and are now prioritizing scaling. The key 672 00:39:08,280 --> 00:39:13,239 Speaker 2: areas of focus for organizations now include managing multiple AI models, 673 00:39:13,680 --> 00:39:18,800 Speaker 2: as well as thinking about specific use cases and desired outcomes. However, 674 00:39:18,880 --> 00:39:22,000 Speaker 2: this scale is difficult for companies to do on their own. 675 00:39:22,600 --> 00:39:27,560 Speaker 2: To unlock the real potential of generative AI in transforming experiences, 676 00:39:27,680 --> 00:39:32,480 Speaker 2: they'll require the scaling capabilities made possible by partnerships like 677 00:39:32,600 --> 00:39:37,320 Speaker 2: IBM and Salesforce. This conversation showed the promise of teamwork. 678 00:39:38,040 --> 00:39:42,359 Speaker 2: When massive companies combine their brain power to push forward technology, 679 00:39:42,680 --> 00:39:49,239 Speaker 2: their collaborative efforts have the potential to revolutionize industries. One 680 00:39:49,360 --> 00:39:52,279 Speaker 2: quick programming note, we will be taking a little time 681 00:39:52,320 --> 00:39:55,120 Speaker 2: off and will be returning in just a few weeks 682 00:39:55,640 --> 00:39:59,400 Speaker 2: with a new episode. Smart Talks with IBM is produced 683 00:39:59,400 --> 00:40:04,759 Speaker 2: by mattro Joey Fishground, David Jaw and Jacob Goldstein. We're 684 00:40:04,880 --> 00:40:08,360 Speaker 2: edited by Lydia Jane Kott. Our engineers are Jason Gambrel, 685 00:40:08,760 --> 00:40:14,560 Speaker 2: Sarah Bruguier and Ben Holliday. Theme song by Gramoscope. Special 686 00:40:14,560 --> 00:40:18,080 Speaker 2: thanks to Andy Kelly, Kathy Callahan and the eight Bar 687 00:40:18,239 --> 00:40:22,080 Speaker 2: and IBM teams, as well as the Pushkin marketing team. 688 00:40:22,440 --> 00:40:25,440 Speaker 2: Smart Talks with IBM is a production of Pushkin Industries 689 00:40:25,680 --> 00:40:30,160 Speaker 2: and Ruby Studio at iHeartMedia. To find more Pushkin podcasts, 690 00:40:30,480 --> 00:40:34,799 Speaker 2: listen on the iHeartRadio app, Apple Podcasts, or wherever you 691 00:40:34,960 --> 00:40:39,440 Speaker 2: listen to podcasts. I'm Malcolm Gladwell. This is a paid 692 00:40:39,480 --> 00:40:47,520 Speaker 2: advertisement from IBM.