1 00:00:00,040 --> 00:00:07,880 We're discussing AI startups today on CXO Talk episode 871 in 2 00:00:07,880 --> 00:00:13,360 a conversation with Arvind Jane, the founder and CEO of Glean. 3 00:00:13,480 --> 00:00:18,040 He was a distinguished engineer at Google before starting and 4 00:00:18,040 --> 00:00:21,560 taking public cybersecurity company Rubric. 5 00:00:22,320 --> 00:00:28,040 Glean has raised $600 million and is valued at almost 5 6 00:00:28,040 --> 00:00:30,560 billion. Glean is an enterprise AI 7 00:00:30,560 --> 00:00:33,480 company. Think of it as Google or Chat 8 00:00:33,480 --> 00:00:36,840 GPD but inside your company. So that's what we do. 9 00:00:37,360 --> 00:00:44,040 And you are a, shall we say, born in AI or AI native company. 10 00:00:44,040 --> 00:00:49,400 So you're the perfect person to help us understand what's going 11 00:00:49,400 --> 00:00:52,360 on with start-ups, AI start-ups today. 12 00:00:52,800 --> 00:00:55,680 Absolutely. We are in fact I think the first 13 00:00:55,680 --> 00:01:00,000 company to bring the transformer technology to the enterprises 14 00:01:00,280 --> 00:01:02,400 and, and and say, yeah, it's a lot of great learnings, you 15 00:01:02,400 --> 00:01:05,519 know, starting as a native Gen. AI company six years back. 16 00:01:05,880 --> 00:01:08,160 Gen. AI company for six years. 17 00:01:08,160 --> 00:01:12,920 That's well before the marketing hype surrounding ChatGPT and 18 00:01:12,920 --> 00:01:14,520 Open AI. That's true. 19 00:01:14,520 --> 00:01:17,280 In fact, the term Gen. AI did not exist at that time. 20 00:01:17,600 --> 00:01:19,960 But the core technology that powers Gen. 21 00:01:19,960 --> 00:01:24,440 AI, the Transformers, they were there and we were using them, 22 00:01:24,680 --> 00:01:29,120 you know, in early 2019 to, you know, understand enterprise 23 00:01:29,120 --> 00:01:33,280 data, knowledge, information at a really deep level using AI and 24 00:01:33,280 --> 00:01:36,880 then making it all searchable for people inside a company. 25 00:01:37,240 --> 00:01:43,120 Can you describe the differences between AI startups and 26 00:01:43,120 --> 00:01:48,200 traditional technology startups? Every company like, you know, 27 00:01:48,200 --> 00:01:51,240 first of all, whether you're a startup or you are a mature 28 00:01:51,240 --> 00:01:55,440 company, I think AI is becoming such a core fundamental tool 29 00:01:55,600 --> 00:01:58,720 that you've got to use in, you know, in products that you 30 00:01:58,720 --> 00:02:01,680 build, because that's, that's the way to sort of stay ahead. 31 00:02:01,680 --> 00:02:03,800 That's the way to actually build new amazing things. 32 00:02:04,080 --> 00:02:06,160 So, so I actually feel like, you know, like this. 33 00:02:06,240 --> 00:02:08,919 My view is that probably all startups that you think of 34 00:02:08,919 --> 00:02:11,960 today, you know, that you can call them all AI startups. 35 00:02:13,040 --> 00:02:16,880 The but, but maybe the other way to look into look into this is 36 00:02:16,880 --> 00:02:19,760 that there are some companies, you know, that are actually 37 00:02:19,760 --> 00:02:22,880 building core foundational technologies. 38 00:02:22,880 --> 00:02:25,200 For example, they're building models, they're building 39 00:02:25,440 --> 00:02:27,480 infrastructure to actually train models. 40 00:02:27,720 --> 00:02:29,600 And maybe, you know, that's what I'm like, you know, you know, I 41 00:02:29,600 --> 00:02:32,600 would say like, you know, one set of companies, but then the 42 00:02:32,600 --> 00:02:36,960 vast majority of AI startups that are getting started today 43 00:02:37,320 --> 00:02:40,320 are companies that are thinking about solving a business 44 00:02:40,320 --> 00:02:46,360 problem, a consumer problem where they feel that AI and you 45 00:02:46,360 --> 00:02:49,840 know, the new reasoning and generation capabilities is going 46 00:02:49,840 --> 00:02:52,880 to play a big role in, in the product that they're going to 47 00:02:52,880 --> 00:02:56,000 build. The but like, you know, like us 48 00:02:56,000 --> 00:02:59,000 living in the AI world, you know, all the, all the startups 49 00:02:59,000 --> 00:03:01,800 that I, you know, engage with, like I'm not, I've actually not 50 00:03:01,800 --> 00:03:04,400 seen a single startup that is not, you know, that that's not 51 00:03:04,400 --> 00:03:06,800 actually making AIA big part of the core tech stack. 52 00:03:07,240 --> 00:03:14,520 As AI has matured, there's become a greater clarity that AI 53 00:03:14,520 --> 00:03:18,960 as you just described must support the solving of some 54 00:03:18,960 --> 00:03:25,120 business problem. For AI startups where the the 55 00:03:25,120 --> 00:03:30,720 centerpiece is that AI technology, what are the 56 00:03:31,160 --> 00:03:35,680 differences from again traditional software companies 57 00:03:35,800 --> 00:03:38,800 where yes, they're solving a business problem, but then but 58 00:03:38,800 --> 00:03:42,640 the technology is very different the underlying foundations. 59 00:03:43,200 --> 00:03:46,520 AI technology, you know, moves very fast, it changes, you know, 60 00:03:46,520 --> 00:03:49,600 at A at a very rapid pace. So when you think about product 61 00:03:49,600 --> 00:03:54,120 development, the the new AI startups, you know, they're 62 00:03:54,120 --> 00:03:56,360 they're able to you know, the first of all, they're very lean. 63 00:03:56,360 --> 00:03:59,760 They they are able to actually do a lot of things, you know, 64 00:03:59,840 --> 00:04:03,320 because software programming, building systems with AI, you 65 00:04:03,320 --> 00:04:06,400 know, these things are becoming, you know, much, much easier than 66 00:04:06,400 --> 00:04:08,560 than you know, than in the pre AI world. 67 00:04:09,360 --> 00:04:14,400 So, so I think, I think the like, again, like, you know, I, 68 00:04:14,400 --> 00:04:17,680 I, I, I struggled with your question because I, you know, 69 00:04:17,680 --> 00:04:21,320 like, like my, from my vantage point, like when you think about 70 00:04:21,320 --> 00:04:25,640 investment that is happening, you know, most of the investment 71 00:04:25,640 --> 00:04:28,840 from venture capital actually is coming into companies that are 72 00:04:28,840 --> 00:04:30,960 making AIA big part of their story. 73 00:04:31,880 --> 00:04:36,880 And, and so I can compare more with like the, let's say that 74 00:04:36,880 --> 00:04:39,320 startups, you know, that we're getting started two years back 75 00:04:39,320 --> 00:04:41,600 versus today. The key difference between them 76 00:04:41,600 --> 00:04:46,440 is, well, like thinking about this new model where you can 77 00:04:46,440 --> 00:04:51,360 truly build products, you know, which are a lot more powerful, 78 00:04:51,360 --> 00:04:53,520 you know, a lot more capable than what would what you would 79 00:04:53,520 --> 00:04:55,840 have even, you know, you know, thought of like, you know, a 80 00:04:55,840 --> 00:04:59,120 couple couple years back. But but but I search like, you 81 00:04:59,120 --> 00:05:01,480 know, I think from a company building perspective, like, you 82 00:05:01,480 --> 00:05:05,200 know, startup is a startup. You know, we are AI company and 83 00:05:05,200 --> 00:05:09,480 like you know, when I compared this with my previous startup, I 84 00:05:09,480 --> 00:05:11,880 would say that like, you know, most of what we do doesn't 85 00:05:11,880 --> 00:05:13,480 change. Like, you know, we, we still 86 00:05:13,480 --> 00:05:15,880 have to think about the core business problem. 87 00:05:15,880 --> 00:05:17,880 We have to figure out how we're going to actually solve that, 88 00:05:17,880 --> 00:05:21,400 how we're going to build great technology and and like, you 89 00:05:21,400 --> 00:05:24,280 know, the teams to to then sort of like, you know, bring that 90 00:05:24,280 --> 00:05:27,120 technology to our customers. So so personally, like I have 91 00:05:27,120 --> 00:05:30,160 this feeling that yes, like, you know, AI is becoming a big part 92 00:05:30,160 --> 00:05:33,600 of our technology stack, but fundamentally how we build and 93 00:05:33,600 --> 00:05:36,640 design companies are not changing except for maybe one 94 00:05:36,640 --> 00:05:37,560 thing like, you know, you do here. 95 00:05:37,560 --> 00:05:40,360 Sometimes companies saying that, well, like, you know, now with 96 00:05:40,360 --> 00:05:44,000 AI, you can actually have a one person company that can generate 97 00:05:44,000 --> 00:05:47,160 a billion dollars because, and then so you see a little bit of 98 00:05:47,160 --> 00:05:50,720 that, like, you know, like, you know, in the new generation of 99 00:05:51,840 --> 00:05:55,120 companies, you are seeing some of these, you know, AI companies 100 00:05:55,440 --> 00:05:58,280 like scaling up revenue at the pace that, you know, the 101 00:05:58,280 --> 00:06:00,280 previous industry of SAS companies couldn't. 102 00:06:00,320 --> 00:06:03,640 Like, you know, we have, we saw some, you know, I, I did these 103 00:06:03,640 --> 00:06:06,600 examples all the time, you know, like start up actually reaching 104 00:06:06,840 --> 00:06:10,200 $20 million, you know, in revenue two months, you know, 105 00:06:10,280 --> 00:06:13,480 like after it's got started or a company that's, you know, at 100 106 00:06:13,480 --> 00:06:16,240 million rate, like, you know, within the first year, these 107 00:06:16,240 --> 00:06:19,040 things are not possible. Like in, you know, in many ways, 108 00:06:19,040 --> 00:06:24,160 you know, in the Pai world, But given like you know how fast, 109 00:06:24,160 --> 00:06:26,920 like you know, you can actually build products and, and how 110 00:06:26,920 --> 00:06:29,880 different your products can be compared to the, you know, the 111 00:06:29,880 --> 00:06:33,280 current products in the market. It's allowing people to actually 112 00:06:33,280 --> 00:06:37,440 just fundamentally scale at a different pace and level than 113 00:06:37,440 --> 00:06:39,600 than the previous generation of start-ups. 114 00:06:40,080 --> 00:06:44,760 I just want everybody to know that you can ask your questions 115 00:06:44,880 --> 00:06:51,160 right now on Twitter. There's a tweet chat, X on XI 116 00:06:51,160 --> 00:06:55,560 should say there's a tweet chat and X chat taking place. 117 00:06:56,120 --> 00:06:59,760 Use the hashtag Cxotalk if you want to ask your questions. 118 00:07:00,040 --> 00:07:04,800 If you're watching on LinkedIn, just O your questions into the 119 00:07:04,800 --> 00:07:07,400 chat. Arvind, you raise a very 120 00:07:07,400 --> 00:07:13,520 interesting point just now, how startups, AI startups can 121 00:07:13,880 --> 00:07:17,440 generate so much revenue with relatively few people. 122 00:07:17,840 --> 00:07:21,960 What is it about AI, this technology and the nature of the 123 00:07:21,960 --> 00:07:26,240 startups that are happening around this to enable this 124 00:07:26,240 --> 00:07:29,440 phenomenon? So first you can actually build 125 00:07:29,440 --> 00:07:36,280 products that are very different in their capabilities. 126 00:07:36,960 --> 00:07:40,280 Then like, you know, then things that are out there, when you 127 00:07:40,280 --> 00:07:45,000 think about, for example, software development, there are 128 00:07:45,000 --> 00:07:49,080 companies, you know, that are offering new development 129 00:07:49,080 --> 00:07:53,320 environments, which allows developers to go five times 130 00:07:53,320 --> 00:07:55,240 faster than what they could before. 131 00:07:55,600 --> 00:08:00,920 And, and so there's a, there is such a big leap in, in the 132 00:08:00,920 --> 00:08:03,320 capabilities, you know, that these, you know, that these 133 00:08:03,320 --> 00:08:06,040 products bring, which creates that instant, you know, 134 00:08:06,040 --> 00:08:07,800 excitement in the market for them. 135 00:08:08,040 --> 00:08:10,640 Like, you know, like, you know, these, these new coding tools, 136 00:08:10,880 --> 00:08:13,560 like, you know, even even for us, like, you know, as a native 137 00:08:13,560 --> 00:08:15,880 Gen. AI company, like we, we see it 138 00:08:15,880 --> 00:08:17,360 everyday. Like, you know, people just 139 00:08:17,360 --> 00:08:19,280 like, you know, they, they want to use these tools. 140 00:08:19,280 --> 00:08:21,440 Like, you know, there is a, like, you know, there's, there 141 00:08:21,440 --> 00:08:26,000 is a big, you know, demand, you know, from the ground up, like 142 00:08:26,000 --> 00:08:27,920 there's love for these new AI products. 143 00:08:28,600 --> 00:08:30,920 That is, that's incredible. And that is what's allowing them 144 00:08:30,920 --> 00:08:34,320 to, to actually like launch a product in the market, like, you 145 00:08:34,320 --> 00:08:37,720 know, have a PSG motion and, and like, you know, there you go. 146 00:08:37,720 --> 00:08:39,880 Like, you know, people come in, they want to use these products. 147 00:08:40,440 --> 00:08:44,200 So it's, it's, I think largely it's driven by these brand new 148 00:08:44,200 --> 00:08:48,360 amazing capabilities and software doing things that 149 00:08:48,360 --> 00:08:50,560 people didn't expect, you know, it could do. 150 00:08:50,840 --> 00:08:52,440 So I think I think that's the biggest part. 151 00:08:52,440 --> 00:08:54,600 Like that's what plays that excitement and then instant 152 00:08:54,600 --> 00:08:58,000 demand. But the other thing which which 153 00:08:58,000 --> 00:09:01,200 is actually allowing these companies to have this success, 154 00:09:01,200 --> 00:09:05,200 you know, you know, in a short duration of time is that, you 155 00:09:05,200 --> 00:09:08,400 know, the software development itself has, you know, gotten 156 00:09:08,400 --> 00:09:11,200 very accelerated. When you use these new tools, 157 00:09:11,920 --> 00:09:16,640 you can build products which are a lot more capable than, you 158 00:09:16,640 --> 00:09:18,760 know, current products that we have in the market. 159 00:09:19,040 --> 00:09:21,520 And, and you can do that in a short period of time. 160 00:09:21,560 --> 00:09:24,160 Like you actually put something amazing, you know, in a month, 161 00:09:24,400 --> 00:09:26,800 you know, in two months, you know, something that used to 162 00:09:26,800 --> 00:09:28,760 take you like a year or two years before. 163 00:09:29,400 --> 00:09:31,960 So that's that's the other factor that is actually, you 164 00:09:31,960 --> 00:09:36,240 know, shrinking these these cycles for these companies. 165 00:09:36,600 --> 00:09:41,720 So you have this combination of tremendous interest and market 166 00:09:41,720 --> 00:09:49,760 demand combined with significant developer productivity increases 167 00:09:50,040 --> 00:09:52,720 as a result of the tools. Is that a correct way of saying 168 00:09:52,720 --> 00:09:54,440 it? That's right, yes. 169 00:09:54,800 --> 00:09:59,000 We have an interesting question coming from LinkedIn and this is 170 00:09:59,000 --> 00:10:06,000 from Santosh Sirivo, who says who is actually using and paying 171 00:10:06,000 --> 00:10:11,200 for AI services. AI has captured everybody's 172 00:10:11,200 --> 00:10:13,760 imagination. There is no enterprise in the 173 00:10:13,760 --> 00:10:18,520 world today that feels that well, like, you know, we can, we 174 00:10:18,520 --> 00:10:22,600 can, we can look into AI like, you know, a few years from now. 175 00:10:23,200 --> 00:10:28,880 Everybody knows that they have to act, they have to embrace AI 176 00:10:29,240 --> 00:10:33,320 today if, if you want to be, you know, if you want to stay 177 00:10:33,320 --> 00:10:36,680 relevant. And so, so starting from that, 178 00:10:36,680 --> 00:10:41,000 starting from that, like high level desire to actually, you 179 00:10:41,000 --> 00:10:42,880 know, bring AI into your enterprise. 180 00:10:43,360 --> 00:10:46,800 Now what we're seeing is that there are two different ways our 181 00:10:46,800 --> 00:10:52,440 people are bringing AI into the company. 1 is the CIO, you know 182 00:10:52,440 --> 00:10:57,200 that, you know, CIOs like, you know, obviously hold the, the, 183 00:10:57,920 --> 00:11:01,000 the responsibility to bring the right set of technologies, you 184 00:11:01,000 --> 00:11:03,800 know, to, to the entire working population of your, of your 185 00:11:03,800 --> 00:11:06,240 enterprise. So they are taking the 186 00:11:06,240 --> 00:11:11,800 leadership in terms of bringing, you know, bringing godly, 187 00:11:11,800 --> 00:11:16,160 applicable, useful products that are AI based to their 188 00:11:16,160 --> 00:11:18,120 enterprise. And that's for, for us, for 189 00:11:18,120 --> 00:11:21,040 example, at Glean, you know, you know, they are our primary 190 00:11:21,040 --> 00:11:25,280 personas too, because Glean as a tool, you know, it's, it's a 191 00:11:25,280 --> 00:11:28,240 knowledge access tool. Like people go to glean, they 192 00:11:28,240 --> 00:11:30,920 ask questions, you know, we quickly answer those questions 193 00:11:30,920 --> 00:11:33,840 for them using all of the enterprise context and data and 194 00:11:33,840 --> 00:11:35,680 knowledge. And so that's the tool. 195 00:11:35,680 --> 00:11:38,240 That's actually every knowledge worker, like, you know, whether 196 00:11:38,240 --> 00:11:42,440 you're engineer, the support person, somebody in HR or IT or 197 00:11:42,440 --> 00:11:44,520 sales, all of you have the need for that. 198 00:11:44,520 --> 00:11:47,800 All of us have questions, all of us have tasks that we think AI 199 00:11:47,800 --> 00:11:51,440 can do for us now. So it's a broad tool and CI OS 200 00:11:51,440 --> 00:11:54,640 are the ones who actually typically like, you know, will, 201 00:11:55,600 --> 00:11:57,840 you know, purchase a a company by two like that. 202 00:11:58,640 --> 00:12:03,120 But then you also have every individual functional leader. 203 00:12:03,960 --> 00:12:07,480 You know, if you are head of support, if you are the CTO and 204 00:12:07,680 --> 00:12:10,200 you're trying to actually make sure that you know, your 205 00:12:10,200 --> 00:12:12,680 engineers, your developers are productive with AI. 206 00:12:12,680 --> 00:12:16,480 If you're, if you are a sales person, like you know, the sales 207 00:12:16,480 --> 00:12:20,360 leader and you want to make sure that you're using modern ways 208 00:12:20,360 --> 00:12:24,560 to, to prospect, to reach out to your customers, to actually have 209 00:12:24,560 --> 00:12:27,840 powerful engagements with them. Function by function. 210 00:12:27,840 --> 00:12:31,360 You're seeing every functional leader in the enterprises 211 00:12:31,360 --> 00:12:34,600 looking into and evaluating AI tools and bringing them on 212 00:12:34,600 --> 00:12:36,520 board. So this is a, this is a very, 213 00:12:36,520 --> 00:12:39,560 very broad phenomena, industry wide vertical, wide 214 00:12:39,800 --> 00:12:42,200 geographical. Like, you know, I've been 215 00:12:42,200 --> 00:12:44,440 travelling a lot across the world. 216 00:12:44,640 --> 00:12:47,480 I don't, I don't, I don't see any company that is not paying 217 00:12:47,480 --> 00:12:50,200 attention to AI, any country that's not paying attention to 218 00:12:50,280 --> 00:12:52,400 AI either. So this is, you know, like is, 219 00:12:52,400 --> 00:12:55,440 is everyone really is the answer like you know, is bringing AI 220 00:12:55,440 --> 00:12:57,160 into their into their enterprises. 221 00:12:57,560 --> 00:13:02,600 Subscribe to our newsletter Join our community Go to cxotalk.com 222 00:13:02,880 --> 00:13:05,800 Subscribe to the newsletter. Check it out. 223 00:13:06,040 --> 00:13:11,760 So it's following a traditional enterprise software, the 224 00:13:11,760 --> 00:13:15,320 purchase process in a way because as you just described, 225 00:13:15,520 --> 00:13:20,440 you've got IT and the CIO and at the same time you have a 226 00:13:20,720 --> 00:13:24,480 functional line of business leaders, HR, whatever, 227 00:13:24,480 --> 00:13:28,880 marketing, whatever it might be. Both of these groups are looking 228 00:13:28,880 --> 00:13:31,600 at AI products. And I'm assuming this comes 229 00:13:31,600 --> 00:13:35,120 right back to what you were saying earlier, which is the 230 00:13:35,120 --> 00:13:37,680 core issue. It's not actually the 231 00:13:37,680 --> 00:13:40,320 technology, it's the business problem that's being solved. 232 00:13:41,920 --> 00:13:47,920 We have a very interesting question from LinkedIn and keep 233 00:13:47,920 --> 00:13:51,360 your questions coming in. We have some questions on 234 00:13:51,360 --> 00:13:53,760 Twitter as well, and we're going to get to all of these. 235 00:13:54,400 --> 00:14:00,800 Risha Varshney asks, how do your offerings address the AI ethics 236 00:14:01,280 --> 00:14:07,200 jailbreak, prompt leakage and AI transparency slash 237 00:14:07,200 --> 00:14:11,400 interpretability risks? And I'll just mention that Risha 238 00:14:11,400 --> 00:14:16,680 is Senior Director of Risk Systems Development at Freddie 239 00:14:16,680 --> 00:14:18,920 Mac. And of course, Financial 240 00:14:18,920 --> 00:14:21,560 Services is keenly interested in these topics. 241 00:14:21,880 --> 00:14:25,280 So Lean is an enterprise AI solution. 242 00:14:25,680 --> 00:14:29,400 We work with the largest enterprises out there in 243 00:14:29,400 --> 00:14:32,480 different industry verticals. We work, we work a lot with 244 00:14:32,480 --> 00:14:35,600 financial services and these risks are all very real. 245 00:14:35,600 --> 00:14:38,800 Like, you know, I think with AI, the, you know, it's a very 246 00:14:38,800 --> 00:14:43,800 powerful technology And, and it's also like a, it's, it's, 247 00:14:43,840 --> 00:14:46,680 it's, you know, it's grounds for, you know, innovation and 248 00:14:46,680 --> 00:14:48,480 like folks who are trying to actually create security 249 00:14:48,480 --> 00:14:52,720 problems and, and you have to be very, very careful in terms of 250 00:14:52,720 --> 00:14:55,960 rolling the technologies in the right way, in a secure way. 251 00:14:56,280 --> 00:15:00,000 And also like, you know, in a way which basically ensures 252 00:15:00,000 --> 00:15:05,240 that, you know, the technology is unbiased and, and, and it's, 253 00:15:05,280 --> 00:15:08,480 you know, it's doing stuff that you think is, you know, safe, 254 00:15:08,480 --> 00:15:10,880 you know, for your, you know, in your context, in your 255 00:15:10,880 --> 00:15:12,840 enterprise. So some, I'll give you some 256 00:15:12,840 --> 00:15:14,920 examples of like, you know, things, you know, problems that 257 00:15:14,920 --> 00:15:17,040 you have to solve on the security front. 258 00:15:18,200 --> 00:15:21,120 One of the key, you know, things that AI is that if you're going 259 00:15:21,120 --> 00:15:25,520 to make it useful for your enterprise, your, for your 260 00:15:25,520 --> 00:15:29,040 organization, you know, you have to take these language models 261 00:15:29,040 --> 00:15:32,280 that are built using the public web data. 262 00:15:32,440 --> 00:15:35,680 They don't really know much about your enterprise and your, 263 00:15:35,680 --> 00:15:39,600 your, you know, you know, your information, your data. 264 00:15:40,560 --> 00:15:42,960 And so somehow you have to figure out how you're going to 265 00:15:42,960 --> 00:15:46,960 actually connect that enterprise context with the power of these 266 00:15:46,960 --> 00:15:50,800 language models and, and do it in a safe and secure way. 267 00:15:50,960 --> 00:15:53,920 For example, if you train the model and you connected all of 268 00:15:53,920 --> 00:15:57,840 your enterprise data and knowledge, you know, with that 269 00:15:57,840 --> 00:16:01,040 model and now anybody can go inside your company and ask 270 00:16:01,040 --> 00:16:04,680 questions, well, it's going to leak a lot of sensitive data to 271 00:16:04,680 --> 00:16:07,240 people who should not have actually, you know, have access 272 00:16:07,240 --> 00:16:09,840 to that information. Because enterprise data and 273 00:16:09,840 --> 00:16:14,680 knowledge is very, you know, it's it's private in nature, 274 00:16:14,680 --> 00:16:16,760 like, you know, like a given document, like, you know, only a 275 00:16:16,760 --> 00:16:19,040 few people may have access to it inside the company. 276 00:16:19,240 --> 00:16:22,560 So you can't take like, you know, our content wholesale, 277 00:16:22,760 --> 00:16:24,760 like you know, in your enterprise and just train or 278 00:16:24,760 --> 00:16:28,400 build any I system with it. Any AI system that you build, it 279 00:16:28,400 --> 00:16:32,240 has to understand who's using, you know, that that particular 280 00:16:32,240 --> 00:16:35,600 system or software. And how do you make sure that 281 00:16:35,800 --> 00:16:39,280 you know, AI only uses information that you know, this 282 00:16:39,280 --> 00:16:42,600 person, you know, is entitled to is, you know, allowed to use 283 00:16:42,600 --> 00:16:45,080 inside the company and create those safe and secure AI 284 00:16:45,080 --> 00:16:46,520 experiences. So that's one of the problems 285 00:16:46,520 --> 00:16:50,240 that we solve. Like in Glean, any AI usage, you 286 00:16:50,240 --> 00:16:55,520 know, any agent that you build on Glean, you have to sort of 287 00:16:55,720 --> 00:16:59,000 use that as an employee and you have to be signed in. 288 00:16:59,240 --> 00:17:02,840 And then we will actually ensure that whatever we do for you is 289 00:17:02,840 --> 00:17:05,520 done with knowledge and information that you could 290 00:17:05,520 --> 00:17:08,200 access to. I would like you to join the 291 00:17:08,400 --> 00:17:14,960 Cxotalk community, so go to cxotalk.com and sign up for our 292 00:17:14,960 --> 00:17:18,800 mailing list so we can notify you about upcoming shows because 293 00:17:18,800 --> 00:17:20,960 we have amazing discussions like this. 294 00:17:21,000 --> 00:17:27,720 All right, let's go to Twitter and to Arsalan Khan who says, do 295 00:17:27,720 --> 00:17:33,040 you think AI will become a simple plug and will become 296 00:17:33,080 --> 00:17:39,240 simply plug and play even for non-technical employees in the 297 00:17:39,240 --> 00:17:45,400 enterprise, so the the broader ease of use for the rest of us? 298 00:17:45,600 --> 00:17:48,880 Well, absolutely. I'm both hopeful and confident, 299 00:17:49,040 --> 00:17:50,680 you know, that that's what's going to happen. 300 00:17:50,880 --> 00:17:54,880 I think the true power of AI will be realized when it's, you 301 00:17:54,880 --> 00:17:58,840 know, so easy to use. You know, you can actually do 302 00:17:58,840 --> 00:18:04,000 things with AI as as a business user, you know, you should not 303 00:18:04,000 --> 00:18:07,200 be required to understand, like you know, how to code, how to 304 00:18:07,200 --> 00:18:10,040 build systems. You know, be an engineer, you 305 00:18:10,040 --> 00:18:12,920 know, you like, you know, as you know, for example, like let's 306 00:18:12,920 --> 00:18:16,680 say you are, you are an, you know, you are an, an employee in 307 00:18:16,680 --> 00:18:19,720 the legal department and you review contracts everyday. 308 00:18:19,960 --> 00:18:22,000 And you know, you know, it's a, it's a process, you know, that 309 00:18:22,000 --> 00:18:25,720 takes you a lot of time, you know, and you should be able to 310 00:18:25,720 --> 00:18:27,720 just, you know, ask AI. You can, you know, like, you 311 00:18:27,720 --> 00:18:30,720 know, and say that like, look, this is, this is, this is how I 312 00:18:30,720 --> 00:18:32,960 review a contract. This is the process that I 313 00:18:32,960 --> 00:18:34,640 follow. And you should be able to just 314 00:18:34,640 --> 00:18:37,760 say that, you know, to an AI system that actually then goes 315 00:18:37,760 --> 00:18:40,240 and identifies, you know, that particular business process for 316 00:18:40,240 --> 00:18:43,680 you, automates it for you. So AI has to truly become 317 00:18:43,680 --> 00:18:46,640 accessible like that with our Asian PIC platform. 318 00:18:46,640 --> 00:18:49,840 That's what we do. You know, we are actually really 319 00:18:49,840 --> 00:18:53,680 thinking about how to make this technology super accessible 320 00:18:53,680 --> 00:18:56,200 like, you know, make, you know, make sure that it doesn't matter 321 00:18:56,200 --> 00:18:58,080 who you are. You could be in HR and, you 322 00:18:58,080 --> 00:19:02,640 know, in finance and legal and you, you, you know, you don't 323 00:19:02,640 --> 00:19:04,840 need to know anything about AI, how it works. 324 00:19:04,840 --> 00:19:06,560 Like, you know what, what are language models? 325 00:19:06,760 --> 00:19:08,440 Like that's, that's not like, you know, that's not what you 326 00:19:08,440 --> 00:19:10,640 need to worry about. You need to just like work with 327 00:19:10,920 --> 00:19:12,760 a smart. You're like working with AI 328 00:19:12,760 --> 00:19:15,000 should be like, you know, working with a really smart 329 00:19:15,000 --> 00:19:18,680 person who you could actually go and get some work to, you know, 330 00:19:18,680 --> 00:19:21,040 you just tell them, you know, like how you know, like how this 331 00:19:21,040 --> 00:19:23,600 work needs to happen and then AI just makes it happen for you. 332 00:19:23,600 --> 00:19:26,440 That's, that's the, that's the model that you're going to see, 333 00:19:27,280 --> 00:19:29,960 you know, with any like the agent, you know, AI agent 334 00:19:29,960 --> 00:19:32,400 platforms that are going to succeed in the market are going 335 00:19:32,400 --> 00:19:34,520 to be of that nature. Like, you know, you have to 336 00:19:34,720 --> 00:19:37,720 elevate, you know, like the, you know, the capabilities of these 337 00:19:37,720 --> 00:19:40,800 systems, you have to make them more accessible to non-technical 338 00:19:40,800 --> 00:19:42,800 users. And the other thing that that 339 00:19:42,800 --> 00:19:45,680 that also add to it like to go one more step, you know, beyond 340 00:19:45,680 --> 00:19:49,480 that people are not seeking, people are not seeking help, you 341 00:19:49,480 --> 00:19:52,080 know, from AI as much as you would expect. 342 00:19:53,040 --> 00:19:55,360 Like, you know, we all have habits, you know, we do things 343 00:19:55,360 --> 00:19:57,360 the way we do. Like, you know, like there's a 344 00:19:57,360 --> 00:20:00,400 lot of inertia in terms of like, you know, thinking about like, 345 00:20:00,400 --> 00:20:02,560 well, should I do this task differently? 346 00:20:02,800 --> 00:20:03,840 Like people don't think that way. 347 00:20:03,840 --> 00:20:06,080 Like, you know, you, you don't have time often. 348 00:20:06,080 --> 00:20:07,680 Like, you know, most of the times you yeah, you don't even 349 00:20:07,680 --> 00:20:09,680 have time to think about, well, should I change? 350 00:20:09,680 --> 00:20:12,360 Like, you know, the way I work and so. 351 00:20:12,840 --> 00:20:16,800 It's not only that AI has to be easy and you need to be able to 352 00:20:17,000 --> 00:20:19,280 summon it and, you know, get it to do work for you. 353 00:20:19,720 --> 00:20:21,320 You know, AI also has to follow you. 354 00:20:21,320 --> 00:20:23,920 And it has to come to you and say that, look, yeah, I'm, I'm 355 00:20:23,920 --> 00:20:27,440 observing that you're doing this work every day and spending, you 356 00:20:27,440 --> 00:20:29,080 know, two hours trying to get this work done. 357 00:20:29,080 --> 00:20:32,080 And I can help you with it. So, you know, it has to come to 358 00:20:32,080 --> 00:20:34,560 you. And and that is when, like, you 359 00:20:34,560 --> 00:20:37,320 know, you really, you know, see, people will embrace the 360 00:20:37,320 --> 00:20:40,560 technology at a large scale. I'm looking forward to that day 361 00:20:40,560 --> 00:20:44,240 happening because I can tell you I use large language models 362 00:20:44,320 --> 00:20:48,760 every day and I use multiple language models and different 363 00:20:48,760 --> 00:20:51,920 modes, research, non research and so forth. 364 00:20:52,520 --> 00:20:56,360 And so much depends on the model you're using. 365 00:20:56,560 --> 00:21:00,680 It seems the time of day, the phases of the moon, how you 366 00:21:00,680 --> 00:21:03,080 construct your prompt, and then the whole thing's just a pain in 367 00:21:03,080 --> 00:21:04,760 the butt. Yeah, it is hard. 368 00:21:04,760 --> 00:21:07,480 Like it's not easy today. And then by the way, you are 369 00:21:07,480 --> 00:21:11,120 using like what I would say probably one of the most, you 370 00:21:11,120 --> 00:21:13,320 know, one of the most accessible tools, like, you know, you're 371 00:21:13,320 --> 00:21:16,560 just talking to a system like in natural language. 372 00:21:16,720 --> 00:21:19,480 Yes, you have a few knobs to select hidden there. 373 00:21:20,320 --> 00:21:22,120 So it's, but it's going to get easier like, you know, and 374 00:21:22,120 --> 00:21:24,880 that's what like our job is, you know, at clean, like one of the 375 00:21:24,880 --> 00:21:28,360 things that we do is so we're not an LLM company, We're not, 376 00:21:28,720 --> 00:21:31,280 you know, building and training these foundation models. 377 00:21:31,480 --> 00:21:35,520 But we are like, our goal is to see that you as a business user, 378 00:21:35,680 --> 00:21:37,960 how do we make sure that all this innovation that's happening 379 00:21:37,960 --> 00:21:42,080 in the industry, how do we make it more accessible to you and, 380 00:21:42,080 --> 00:21:45,520 and make things easy, you know, as seamless as like, again, 381 00:21:45,520 --> 00:21:48,920 like, you know, the working model for me with AI is, well, 382 00:21:49,160 --> 00:21:52,360 AI is like a smart human that's been in your company from day 383 00:21:52,360 --> 00:21:54,000 one. They know everything about your 384 00:21:54,000 --> 00:21:55,760 company. They know all the people, 385 00:21:55,880 --> 00:21:59,360 they've read all the documents, all the, all they've been part 386 00:21:59,360 --> 00:22:01,800 of every single meeting and they're ready to now help you 387 00:22:01,920 --> 00:22:04,120 24/7. Just ask them what you want and 388 00:22:04,120 --> 00:22:05,920 they do it for you. Like that's, that's the, that's 389 00:22:05,920 --> 00:22:08,800 the right model for when you know AI is going to deliver true 390 00:22:08,800 --> 00:22:12,560 value in enterprise. Greg Walter says glean looks to 391 00:22:12,560 --> 00:22:17,920 be the glue or connector between disparate databases. 392 00:22:18,440 --> 00:22:22,360 Do you see a future where this function is no longer needed? 393 00:22:22,800 --> 00:22:26,480 Enterprise environments are so complex today. 394 00:22:26,840 --> 00:22:29,400 You know, any large enterprise, you'll get like, you know, 395 00:22:29,400 --> 00:22:33,520 thousands of systems, you know, the data depositories, 396 00:22:33,920 --> 00:22:39,040 databases, you know, unstructured, you know, data 397 00:22:39,040 --> 00:22:43,480 depositories, you know, documents and the like. 398 00:22:43,480 --> 00:22:47,360 The true magic of AI happens when you have access to all of 399 00:22:47,360 --> 00:22:50,360 that information across all of these different systems. 400 00:22:51,840 --> 00:22:54,640 Maybe like, I think things are actually going to get simpler in 401 00:22:54,640 --> 00:22:57,560 the sense that, you know, AI is going to get smarter and 402 00:22:57,560 --> 00:23:00,520 smarter, like in terms of, you know, being able to connect with 403 00:23:00,520 --> 00:23:02,320 all of those different systems over time. 404 00:23:02,680 --> 00:23:06,000 And, and a lot of things that we do today at lean, like, you 405 00:23:06,000 --> 00:23:08,480 know, which we have to actually do a lot of hard work for to 406 00:23:08,480 --> 00:23:10,760 actually connect with these different enterprise systems. 407 00:23:11,000 --> 00:23:13,640 Like we do expect, you know, it to get easier over time. 408 00:23:14,840 --> 00:23:18,280 So yes, like, you know, some like, I think that's the, you 409 00:23:18,280 --> 00:23:20,680 know, if we're, if we, if the answer is that like, you know, 410 00:23:20,680 --> 00:23:23,200 the complexities are not going to go away, you know, then I 411 00:23:23,200 --> 00:23:29,360 think he is not doing his job. Dave Brace on LinkedIn asks, is 412 00:23:29,360 --> 00:23:34,920 it likely that non deterministic agentic systems can honestly 413 00:23:34,920 --> 00:23:39,560 prove that they are truly trustworthy in the enterprise 414 00:23:39,880 --> 00:23:43,880 and that business leaders can trust products like Glean to 415 00:23:43,880 --> 00:23:48,160 make thousands of business decisions every day? 416 00:23:48,760 --> 00:23:53,400 AI is non deterministic. It can also be wrong. 417 00:23:53,880 --> 00:23:59,400 It can hallucinate and to some degree, like, you know, humans 418 00:23:59,400 --> 00:24:01,360 are also like that, like, you know, if you, you know, if you 419 00:24:01,360 --> 00:24:04,120 have questions and you go and ask somebody, like sometimes, 420 00:24:04,120 --> 00:24:06,240 you know, they don't have the full context, they're going to 421 00:24:06,240 --> 00:24:09,720 give you an answer and it may be wrong or it may be incomplete. 422 00:24:10,040 --> 00:24:13,840 And, and so I think the, it's, that's, that's, that's the 423 00:24:13,840 --> 00:24:17,360 fundamental thing to remember that, you know, these AI systems 424 00:24:17,680 --> 00:24:21,280 are actually more like humans, you know, and, and less like 425 00:24:21,280 --> 00:24:25,480 machines, you know, that they used to and, and now you have to 426 00:24:25,480 --> 00:24:28,480 figure out like, well, how do I use this technology? 427 00:24:28,480 --> 00:24:31,080 Like this is not perfect, is going to make mistakes. 428 00:24:31,320 --> 00:24:34,120 So how do I trust it, you know, with, you know, my mission 429 00:24:34,120 --> 00:24:37,120 critical processes where precision is actually, you know, 430 00:24:37,120 --> 00:24:40,760 absolutely required. And so, so there are a few 431 00:24:40,760 --> 00:24:43,080 different strategies that I would say that like, you know, 432 00:24:43,240 --> 00:24:45,880 you know, as an enterprise, you can think about there's a lot of 433 00:24:45,880 --> 00:24:48,800 work where you don't need precision, you know, you know, 434 00:24:48,800 --> 00:24:51,280 work where you need creativity. And that's where like this 435 00:24:51,280 --> 00:24:53,840 technology is already really, really well suited to do. 436 00:24:54,960 --> 00:24:58,120 But then when you think about like your task with that require 437 00:24:58,120 --> 00:25:02,200 precision, AI can be actually used in a few different ways. 1 438 00:25:02,200 --> 00:25:05,120 is that, let's say that there's a business process and now 439 00:25:05,120 --> 00:25:07,840 you're going to agentify, agentify that business process, 440 00:25:08,720 --> 00:25:12,440 you're going to automate it. You're going to ask AI to, to 441 00:25:12,440 --> 00:25:15,600 understand that business process and come up with a plan with the 442 00:25:15,600 --> 00:25:20,600 workflow to, to actually execute that business process from now 443 00:25:20,600 --> 00:25:22,840 on. And so when you, when you use AI 444 00:25:22,840 --> 00:25:25,480 in this fashion, like, you know, you will go and you know, ask AI 445 00:25:25,480 --> 00:25:28,920 to actually build that agent for you and like, maybe to make 446 00:25:28,920 --> 00:25:31,840 mistakes and but you know, you should, you should be there. 447 00:25:32,040 --> 00:25:34,520 Go and supervise it like, you know, go and like ask it to 448 00:25:34,520 --> 00:25:38,160 tweak, you know, it's work, you know, go manually fix, edit it. 449 00:25:38,320 --> 00:25:42,000 And ultimately in that collaboration with AI, you 450 00:25:42,000 --> 00:25:44,960 actually encode and build that agent. 451 00:25:45,000 --> 00:25:47,280 You know that workflow. But now this workflow is 452 00:25:47,280 --> 00:25:49,520 deterministic. Like you, you, you put some 453 00:25:49,520 --> 00:25:52,400 investment in it. AI helped you like, you know, 454 00:25:52,400 --> 00:25:54,840 build this workload very quickly, but you were totally in 455 00:25:54,840 --> 00:25:57,840 control, you were monitoring it. You've gotten into a coordinate 456 00:25:57,840 --> 00:26:00,960 into a place where now, now as I said, it's deterministic. 457 00:26:01,120 --> 00:26:04,240 And now this business process can actually run and it is OK. 458 00:26:04,240 --> 00:26:06,040 Like you know, you don't, you don't need full automation. 459 00:26:06,040 --> 00:26:09,400 Like you can actually work with AI, you know, and spend like you 460 00:26:09,400 --> 00:26:13,040 know that, you know, initial time like an hour or two, but 461 00:26:13,040 --> 00:26:15,120 now you're going to have like, you know, this automation for 462 00:26:15,120 --> 00:26:17,720 years because and it's no longer non deterministic. 463 00:26:17,720 --> 00:26:20,600 So, so you don't think about it like the technology allows you 464 00:26:20,960 --> 00:26:25,040 to to know great things and you don't have to rely on AI to 465 00:26:25,040 --> 00:26:29,160 actually make complex decisions. You know, behind the scenes you 466 00:26:29,160 --> 00:26:32,200 can actually work with it that initially and and build 467 00:26:32,200 --> 00:26:33,800 deterministic systems with it as well. 468 00:26:34,360 --> 00:26:39,760 So you're describing essentially the role of that phrase. 469 00:26:39,760 --> 00:26:43,480 We often hear the human in the loop and what's the appropriate 470 00:26:43,480 --> 00:26:50,680 relationship between the the the person and the AI system that at 471 00:26:50,680 --> 00:26:55,240 this stage of development is a tool rather than I'm looking at 472 00:26:55,280 --> 00:27:00,240 LinkedIn and Greg, Greg Walters says the true magic of AI 473 00:27:00,240 --> 00:27:02,400 replacing old standard applications. 474 00:27:02,840 --> 00:27:05,120 Let's take some examples. You know, we initially were 475 00:27:05,120 --> 00:27:08,840 talking about a legal person that reviews contracts. 476 00:27:09,360 --> 00:27:14,080 Now, if you get 100 page, you know, you know, agreement, 477 00:27:14,080 --> 00:27:17,680 customer agreement, and it's going to take you like, you 478 00:27:17,680 --> 00:27:21,280 know, a week or two weeks to actually go and review this and 479 00:27:21,280 --> 00:27:24,520 make sure that you know, all the terms and conditions, you know, 480 00:27:24,520 --> 00:27:27,880 meet, you know, your enterprises requirements and you're to 481 00:27:27,880 --> 00:27:31,440 redline this document. And and you didn't say that 482 00:27:31,440 --> 00:27:33,520 like, you know, well, I don't trust AI to do this for me. 483 00:27:33,520 --> 00:27:36,240 Like this is pretty sensitive stuff, but well, you know, get 484 00:27:36,240 --> 00:27:39,760 AI to do the first version of the redlining and you know, it's 485 00:27:39,760 --> 00:27:41,520 going to do a great job. Like, you know, if you just tell 486 00:27:41,520 --> 00:27:44,320 it like how you do it, you tell that to AI, it's going to get 487 00:27:44,320 --> 00:27:47,720 90% of the way there. And if it gets 90% of the way 488 00:27:47,720 --> 00:27:51,080 there and now you can actually, you know, like fine tune that 489 00:27:51,080 --> 00:27:54,840 and finish that work, you know, with the context and know how 490 00:27:55,040 --> 00:27:57,640 well, like, you know, that two week task, now it's actually, 491 00:27:57,640 --> 00:27:59,040 you know, a one day task for you. 492 00:27:59,560 --> 00:28:01,520 And that's, that's, that's big. Like, you know, you don't have 493 00:28:01,520 --> 00:28:06,800 to, you don't have to actually like aim for 100% automation and 494 00:28:06,800 --> 00:28:09,280 like, you know, remove yourself from the task completely. 495 00:28:09,280 --> 00:28:11,680 Like, you know, there's, you know, we can, you know, I'm very 496 00:28:11,680 --> 00:28:14,200 happy if I get 90%. You know, it's a big impact to 497 00:28:14,200 --> 00:28:17,480 the business. This is from Gersheron S on 498 00:28:17,480 --> 00:28:23,080 LinkedIn who says which agents are driving the most value in 499 00:28:23,080 --> 00:28:28,280 large enterprise functions. Even better if you have examples 500 00:28:28,280 --> 00:28:33,080 in an industrial AI context. And I should mention that 501 00:28:33,280 --> 00:28:38,560 Gershwan is an AI product manager in metals mining. 502 00:28:38,880 --> 00:28:41,760 So which agents are driving the most value in enterprise 503 00:28:41,760 --> 00:28:46,800 functions? Top use cases for AI today in 504 00:28:46,800 --> 00:28:50,320 the enterprise are the following three. 505 00:28:50,600 --> 00:28:56,160 Number one is general knowledge access and assistance. 506 00:28:56,440 --> 00:29:00,040 So you know, you are a knowledge worker, you, you may be in, you 507 00:29:00,040 --> 00:29:03,320 know, healthcare or financial services or you know, industrial 508 00:29:03,320 --> 00:29:07,280 sector and you have questions, you know, you have questions 509 00:29:07,280 --> 00:29:09,840 that you need answers to. You have information that you 510 00:29:09,840 --> 00:29:14,400 need to do your tasks and, and you use AI to actually, you 511 00:29:14,400 --> 00:29:18,080 know, help you with that. So like tools like ChatGPT or 512 00:29:18,080 --> 00:29:21,160 tools like clean, like inside your company that are just 513 00:29:21,160 --> 00:29:24,280 general purpose, they're not meant for a specific use case. 514 00:29:24,280 --> 00:29:26,480 They're basically knowledge tools like, you know, they help 515 00:29:26,480 --> 00:29:29,800 you, you know, they make knowledge accessible, you know, 516 00:29:29,800 --> 00:29:33,160 from the world, from your enterprise to you so that you 517 00:29:33,160 --> 00:29:35,400 can move faster with, you know, with your tasks. 518 00:29:35,400 --> 00:29:37,240 That's the number one use case for AI today. 519 00:29:37,800 --> 00:29:40,680 And it's not surprising because like this whole revolution was 520 00:29:40,680 --> 00:29:43,520 catalyzed by chat GPD. And so that's, that's the, 521 00:29:43,520 --> 00:29:45,720 that's the application that comes to everybody's mind when 522 00:29:45,720 --> 00:29:49,640 they think about AI. The the second use case is, I 523 00:29:49,640 --> 00:29:52,480 would say for, for software development, which is around 524 00:29:52,480 --> 00:29:55,160 code generation, like developers tasks, like how you build 525 00:29:55,160 --> 00:29:58,360 systems, you know, right, you know, develop technology that's 526 00:29:58,360 --> 00:30:00,400 fundamentally changing with AI. So that's, that's a very 527 00:30:00,400 --> 00:30:02,200 powerful use case. Lot of correction there. 528 00:30:02,480 --> 00:30:04,440 And then the third one I would say is, you know, is around 529 00:30:04,440 --> 00:30:08,840 service. When I say service, I mean like, 530 00:30:08,840 --> 00:30:13,000 you know, I know folks who are actually like taking questions, 531 00:30:13,000 --> 00:30:17,640 taking, you know, complaints tasks from their customers or 532 00:30:17,840 --> 00:30:21,000 their internal employees. So these are like customer 533 00:30:21,000 --> 00:30:24,320 service teams, IT, you know, internal IT help desk, you know, 534 00:30:24,320 --> 00:30:28,440 HR help desk, folks like who are servicing other people's 535 00:30:28,440 --> 00:30:31,280 requests and demands, like, you know, you know, taking, taking 536 00:30:31,280 --> 00:30:34,160 your company's data and knowledge and automating a lot 537 00:30:34,160 --> 00:30:37,160 of that interaction with AI. That's the, that's the big use 538 00:30:37,160 --> 00:30:38,600 case. So functionally those are the, 539 00:30:38,880 --> 00:30:43,560 those are the three top applications and, and, and I 540 00:30:43,560 --> 00:30:46,360 think like across industry verticals like those, that's 541 00:30:46,360 --> 00:30:48,280 what we're seeing. So like I'll give you some 542 00:30:48,280 --> 00:30:52,000 examples on customer service. So we have like these large 543 00:30:52,000 --> 00:30:54,920 telcos, you know, that have like, you know, 50,000 or even 544 00:30:54,920 --> 00:30:57,880 100,000 customer care agents day in day out. 545 00:30:57,880 --> 00:31:01,040 Like, you know, they are getting questions from their customers 546 00:31:01,040 --> 00:31:02,440 that they need to quickly answer. 547 00:31:03,280 --> 00:31:06,120 And you know, if you, if you make those transactions twice as 548 00:31:06,120 --> 00:31:08,480 fast, like, you know, that's, those are hundreds of millions 549 00:31:08,480 --> 00:31:10,360 of dollars of savings, you know, for those teams. 550 00:31:10,560 --> 00:31:13,920 So like that's a very powerful use case for AI today for for 551 00:31:13,920 --> 00:31:17,800 glean, you know, we already talked about software 552 00:31:17,800 --> 00:31:20,160 development, like, you know, like, you know, we have large 553 00:31:20,160 --> 00:31:24,320 enterprises in industrial sector, in retail, you know, 554 00:31:24,320 --> 00:31:30,800 where you are fundamentally changing how, how you use AI to 555 00:31:30,800 --> 00:31:34,920 build systems faster, to test them to, to review, like, you 556 00:31:34,920 --> 00:31:37,280 know, code to troubleshoot. These are some of the, you know, 557 00:31:37,280 --> 00:31:40,800 key use cases as well, like, you know, for, you know, across the 558 00:31:40,800 --> 00:31:43,320 industry. This is, again, from Arsalan 559 00:31:43,320 --> 00:31:47,240 Khan on Twitter. He says AI might be able to 560 00:31:47,240 --> 00:31:52,160 automate standard operating procedures, but what about the 561 00:31:52,200 --> 00:31:56,760 institutional knowledge that resides in people's heads? 562 00:31:57,040 --> 00:32:02,720 And he says, isn't this a significant job security issue? 563 00:32:03,120 --> 00:32:07,720 Today AI like even with Glean, we are able to actually tap into 564 00:32:07,720 --> 00:32:11,720 a lot of their institutional knowledge to then create these 565 00:32:12,000 --> 00:32:15,920 powerful like, you know, GPT like experiences where you can 566 00:32:15,920 --> 00:32:18,880 come and ask questions and you can answer that using, you know, 567 00:32:18,880 --> 00:32:21,320 that institutional knowledge. Then think about institutional 568 00:32:21,320 --> 00:32:25,600 knowledge like it's actually present in a few different form 569 00:32:25,600 --> 00:32:27,880 factors. You know, you have documents 570 00:32:28,120 --> 00:32:31,160 inside your company, you know, where people like, you know, 571 00:32:31,160 --> 00:32:36,800 write stuff, you have ticketing systems, you have like 572 00:32:36,800 --> 00:32:38,800 databases, you know, CRM systems. 573 00:32:38,800 --> 00:32:41,000 So there's a lot of wealth of knowledge inside each one of 574 00:32:41,000 --> 00:32:44,000 these different systems. But then there's also a lot of 575 00:32:44,000 --> 00:32:47,680 knowledge and communication tools like, you know, e-mail or, 576 00:32:47,880 --> 00:32:52,080 or Slack. And there's an, an increasingly 577 00:32:52,240 --> 00:32:54,880 like, you know, enterprises are changing behavior so that they 578 00:32:54,880 --> 00:32:57,760 can capture more and more of that institutional knowledge. 579 00:32:58,240 --> 00:33:01,080 One example of that is that, well, like, you know, if, if two 580 00:33:01,080 --> 00:33:03,440 people are going to actually talk and have a meeting, you 581 00:33:03,440 --> 00:33:06,200 know, record that meeting or at least, you know, like capture 582 00:33:06,200 --> 00:33:09,480 the summary of, you know, that what happened in that meeting 583 00:33:09,680 --> 00:33:12,640 and make it available to AI so that, you know, it can be, you 584 00:33:12,640 --> 00:33:17,680 know, tapped into in the future. So, so I, so I think like so, so 585 00:33:17,680 --> 00:33:19,080 that's, you know, these are all the things that you can do 586 00:33:19,080 --> 00:33:20,760 today. Like, you know, we, we capture 587 00:33:21,000 --> 00:33:24,400 all these forms of institutional knowledge to then actually make 588 00:33:24,400 --> 00:33:26,960 that knowledge work for you as an individual and help you. 589 00:33:28,560 --> 00:33:30,480 And, and like, you know, is driving that. 590 00:33:30,480 --> 00:33:33,760 Like now people are also more motivated to, to actually 591 00:33:33,760 --> 00:33:37,320 capture this information more in our company, for example, like, 592 00:33:37,320 --> 00:33:42,120 you know, we now record like every non, you know, one-on-one 593 00:33:42,120 --> 00:33:43,760 meeting. Like, you know, I could say, you 594 00:33:43,760 --> 00:33:46,480 know, confidential meeting with between the employee and a 595 00:33:46,480 --> 00:33:47,400 manager. We won't. 596 00:33:47,640 --> 00:33:50,480 But like, you know, if it's about, if it's about, you know, 597 00:33:50,480 --> 00:33:52,960 a technical discussion, it's about like getting some work 598 00:33:52,960 --> 00:33:55,680 done, you know, we'll typically record those meetings so that 599 00:33:55,680 --> 00:33:58,440 like all of that data is available to AI in the future to 600 00:33:58,440 --> 00:34:01,640 help us. So that's but but now coming 601 00:34:01,640 --> 00:34:04,920 back to your second question on, well, does it actually, you 602 00:34:04,920 --> 00:34:06,840 know, create an issue with job security? 603 00:34:07,080 --> 00:34:10,679 Like, you know, we, you know, I believe that, you know, the like 604 00:34:10,679 --> 00:34:14,920 as an individual, the strategy for you to make sure that you 605 00:34:14,920 --> 00:34:18,520 stay relevant is you don't go and learn AI like you don't like 606 00:34:18,520 --> 00:34:20,800 this. These tools are amazing and 607 00:34:20,960 --> 00:34:23,920 don't think of AI as something that's going to take a job away 608 00:34:23,920 --> 00:34:26,239 for you from you. I don't think, you know, AI is 609 00:34:26,239 --> 00:34:27,800 that powerful. I don't think it's going to do 610 00:34:27,800 --> 00:34:31,560 it for most people, but you'll certainly like, you know, lose 611 00:34:31,560 --> 00:34:35,360 edge against somebody else who knows how to use these AI tools 612 00:34:35,360 --> 00:34:36,880 and work faster and better than you. 613 00:34:36,880 --> 00:34:41,360 So, so I think that the, the, I, I think that what, what as 614 00:34:41,360 --> 00:34:44,040 individuals, what we need to do is like, you know, like, like, 615 00:34:44,040 --> 00:34:45,679 like, like, like this have been always like that. 616 00:34:45,679 --> 00:34:47,639 Like you know, when new technologies come, folks who 617 00:34:47,639 --> 00:34:51,000 embrace them, like you know who quickly try to learn them, are 618 00:34:51,000 --> 00:34:54,239 the ones who come out ahead. I think this is a very important 619 00:34:54,239 --> 00:34:58,000 point and I certainly give people that same advice that 620 00:34:58,000 --> 00:35:03,480 you, you must be learning how to use AI to make yourself faster, 621 00:35:03,640 --> 00:35:07,920 better, more efficient. And there are so many jobs that 622 00:35:07,920 --> 00:35:12,000 will get displaced. So if you're listening and 623 00:35:12,000 --> 00:35:15,200 you're not doing that, you know, I'm assuming if you're listening 624 00:35:15,200 --> 00:35:19,200 to CXO talk, you are doing that. But so tell, tell the folks you 625 00:35:19,200 --> 00:35:22,560 work with who maybe are not so far on the cutting edge, give 626 00:35:22,560 --> 00:35:24,920 them that, that advice. It's good advice. 627 00:35:26,480 --> 00:35:29,800 So let's talk, come back to startups and we have a question 628 00:35:29,800 --> 00:35:31,880 from Twitter directly on startups. 629 00:35:31,880 --> 00:35:38,000 And this is from Gus Bechdash, who says, Arvind, what advice do 630 00:35:38,000 --> 00:35:41,680 you have for people forming AI ventures? 631 00:35:42,000 --> 00:35:45,400 Many make the mistake of choosing problems that the 632 00:35:45,400 --> 00:35:50,840 platforms will take over, making their companies irrelevant. 633 00:35:51,320 --> 00:35:55,720 I actually don't believe, first of all, that, you know, if you 634 00:35:55,720 --> 00:35:59,520 start a company, if you start to solve the problem that you're 635 00:35:59,520 --> 00:36:02,960 going to fail because of somebody else, because of like, 636 00:36:02,960 --> 00:36:06,600 for example, an incumbent or a large company actually solving 637 00:36:06,600 --> 00:36:08,920 that problem and solving them before you. 638 00:36:09,880 --> 00:36:12,480 You're going to have a firm belief like as an entrepreneur, 639 00:36:12,640 --> 00:36:15,080 that you can solve the problem faster than a large company. 640 00:36:15,080 --> 00:36:17,560 A large company always has lots and lots of things to actually 641 00:36:17,560 --> 00:36:19,200 worry about. Do you know, they're 642 00:36:19,360 --> 00:36:23,520 structurally not designed in a way that they can move faster 643 00:36:23,720 --> 00:36:26,640 than you like as a really, you know, as a nimble early startup? 644 00:36:26,640 --> 00:36:30,360 So so I don't actually agree with that premise of like people 645 00:36:30,360 --> 00:36:32,280 making that mistake. Most of them startups fail 646 00:36:32,280 --> 00:36:34,960 because people give up like, you know, because you lose 647 00:36:34,960 --> 00:36:38,200 confidence in your own idea. You don't have enough conviction 648 00:36:39,080 --> 00:36:41,560 because like, you know, if there's a real problem, well, 649 00:36:41,560 --> 00:36:45,160 you have the right to go and solve it and and you will be and 650 00:36:45,160 --> 00:36:46,920 you know, competition is not something that's going to 651 00:36:46,920 --> 00:36:49,720 matter. So, so with that, with that 652 00:36:49,720 --> 00:36:55,520 said, I will add to what's the right strategy for you to choose 653 00:36:55,520 --> 00:36:58,560 as a founder? Well, like, you know, pick, I 654 00:36:58,560 --> 00:37:02,920 always like to pick problems which are first of all, they're 655 00:37:02,920 --> 00:37:04,680 obvious. Like, you know, like you have to 656 00:37:04,680 --> 00:37:07,960 go and talk to five people. They will not argue with you 657 00:37:07,960 --> 00:37:09,520 that, hey, is this a problem or not? 658 00:37:09,960 --> 00:37:12,280 I think is it all like, if you talk to the first five people 659 00:37:12,280 --> 00:37:15,160 and there and you you don't have that clarity, but you know, from 660 00:37:15,160 --> 00:37:17,480 them, then there's something wrong and you got to like work a 661 00:37:17,480 --> 00:37:20,880 little bit more on it idea. So, so get to that, get to that 662 00:37:20,880 --> 00:37:23,800 level where like, you know, whoever you talk to actually 663 00:37:23,800 --> 00:37:25,840 agree with you that yes, you know, you're solving a real 664 00:37:25,840 --> 00:37:27,200 like, you know, important problem. 665 00:37:28,120 --> 00:37:30,880 And and also like, you know, like I like to work on problems 666 00:37:30,880 --> 00:37:33,520 which have broader impact. So big problems, you know, that 667 00:37:33,520 --> 00:37:35,240 a lot of people are going to have because you know, that's 668 00:37:35,240 --> 00:37:37,320 going to actually create more opportunities for you. 669 00:37:38,400 --> 00:37:41,320 Like even if there are a few other companies that also solve 670 00:37:41,320 --> 00:37:44,200 the same problem where there's a large market that you can tap 671 00:37:44,200 --> 00:37:46,280 into and you'll have your own success, you know, along with 672 00:37:46,280 --> 00:37:47,920 them. And then over time, you know, 673 00:37:48,040 --> 00:37:50,040 like, you know, if you do the best that you're going to win 674 00:37:50,040 --> 00:37:53,600 like over everybody else. And then, and then the, the, the 675 00:37:53,600 --> 00:38:01,240 last thing I would add is don't like come up with an idea where 676 00:38:01,240 --> 00:38:05,280 you feel like, well, like all I need to do is use AI and it's 677 00:38:05,280 --> 00:38:09,040 gonna actually solve this particular task for me. 678 00:38:09,200 --> 00:38:13,800 Like, think about, you know, if it is easy to build, if it is 679 00:38:13,800 --> 00:38:15,880 super easy to build, then everybody else can also build 680 00:38:15,880 --> 00:38:17,160 it. And then you're not really 681 00:38:17,160 --> 00:38:20,840 adding value. So, you know, like AI should be 682 00:38:21,040 --> 00:38:25,160 no more than one of the tools in your tool kit to solve that 683 00:38:25,160 --> 00:38:27,040 problem. But like, you know, make sure 684 00:38:27,040 --> 00:38:28,760 that you know there's something substantial that you're 685 00:38:28,760 --> 00:38:32,040 building. There are so many companies that 686 00:38:32,040 --> 00:38:37,400 are building wrappers around the models. 687 00:38:37,600 --> 00:38:41,520 What about consolidation among these types of companies as the 688 00:38:41,520 --> 00:38:44,840 models advanced their capabilities? 689 00:38:45,200 --> 00:38:49,400 If you are a startup and you are a thin dropper over the core 690 00:38:49,400 --> 00:38:55,360 capabilities of an LLN, well, you'll be irrelevant quite soon. 691 00:38:57,120 --> 00:38:59,240 Yeah, you have to like in the mindset, like I'll tell you what 692 00:38:59,240 --> 00:39:02,600 we do again, you know, you know, we, we know that like, you know, 693 00:39:02,600 --> 00:39:05,480 so the way our product works is that, you know, we actually work 694 00:39:05,480 --> 00:39:08,440 with all the LMS, you know, that are out there in the market. 695 00:39:08,440 --> 00:39:10,880 You know, whether it's, you know, LMS from Open AI or in 696 00:39:10,880 --> 00:39:15,240 Tropic or Google or Meta or, you know, like and all like, you 697 00:39:15,240 --> 00:39:18,760 know, so many LMS from open source, we work with all of 698 00:39:18,760 --> 00:39:21,640 them. And our model for like what we 699 00:39:21,640 --> 00:39:27,160 do is, well, we're going to use all the capabilities that these 700 00:39:27,160 --> 00:39:30,440 LLMS, you know, provide to us and they will actually bring 701 00:39:30,440 --> 00:39:32,160 those capabilities to our customers. 702 00:39:32,440 --> 00:39:35,320 But we'll also be, you know, ensuring that we're building a 703 00:39:35,920 --> 00:39:39,440 very deep technology stack on top of that platform. 704 00:39:39,640 --> 00:39:43,160 And as the LLMS advance, as they actually, you know, add some of 705 00:39:43,160 --> 00:39:45,480 those capabilities that they've built ourselves that actually 706 00:39:45,480 --> 00:39:49,200 throw our, we have to throw what we built like, you know, that 707 00:39:49,200 --> 00:39:53,200 the LLM providers can do already for us and actually keep going 708 00:39:53,200 --> 00:39:55,240 up the value chain. Like, you know, and, and so 709 00:39:55,240 --> 00:39:56,600 that's, that's the model that we choose. 710 00:39:56,600 --> 00:40:00,920 Like, you know, like to get out your space and maximally use 711 00:40:01,160 --> 00:40:04,400 like the innovation that's happening in the industry, but 712 00:40:04,400 --> 00:40:07,000 then you don't build build a significant layer on top of of 713 00:40:07,000 --> 00:40:09,600 that so that you can make that technology accessible to your to 714 00:40:09,600 --> 00:40:13,920 your customers. Given the importance of the 715 00:40:14,160 --> 00:40:18,880 foundation models to your business, how do you manage the 716 00:40:18,880 --> 00:40:24,280 fact that AI is evolving at such a rapid pace? 717 00:40:24,280 --> 00:40:28,920 And how do you balance the stability of your business and 718 00:40:29,000 --> 00:40:33,680 product direction while these underlying capabilities are just 719 00:40:33,680 --> 00:40:36,000 shifting all the time? We have no choice. 720 00:40:36,320 --> 00:40:40,680 You have to like take advantage of the rapid innovation that is 721 00:40:40,680 --> 00:40:43,120 happening in the industry, otherwise you're going to be 722 00:40:43,120 --> 00:40:46,360 left behind and you have to fundamentally change like, you 723 00:40:46,360 --> 00:40:49,440 know, your execution model. Like, you know, we have to like, 724 00:40:49,440 --> 00:40:53,440 you know, like we, when we started, you know, when we 725 00:40:53,440 --> 00:40:57,800 started glean the we sort of very understood that very, you 726 00:40:57,800 --> 00:41:01,080 know, at a fundamental level, like, you know, what 727 00:41:01,080 --> 00:41:04,000 technologies, you know, we had available from open source and 728 00:41:04,000 --> 00:41:06,360 cloud. And, and then you get to sort of 729 00:41:06,360 --> 00:41:08,480 build your road map and you build a, you build a one year 730 00:41:08,480 --> 00:41:11,320 road map. And and that's, that's not no 731 00:41:11,320 --> 00:41:13,240 longer how things work today. Right now. 732 00:41:13,240 --> 00:41:15,200 Technology changes on a monthly basis. 733 00:41:15,440 --> 00:41:17,600 And, and so you have to fundamentally change your 734 00:41:17,600 --> 00:41:19,080 architecture. Like, you know, like one of the 735 00:41:19,080 --> 00:41:21,920 things we change, as an example, is that we don't have the annual 736 00:41:22,320 --> 00:41:24,560 like, you know, or a quarterly planning process. 737 00:41:24,560 --> 00:41:28,280 We switch to a monthly planning process in terms of like, like 738 00:41:28,280 --> 00:41:31,040 how we're going to actually, you know, build our technology 739 00:41:31,200 --> 00:41:33,480 because every month, like, you know, there's new things to look 740 00:41:33,480 --> 00:41:35,240 at and you have to quickly adapt. 741 00:41:35,680 --> 00:41:39,040 The other thing that we, you know, also added is this concept 742 00:41:39,040 --> 00:41:43,800 of the, well, figure out what you're going to throw like every 743 00:41:43,800 --> 00:41:45,960 month, you know, in technology that you've built. 744 00:41:46,240 --> 00:41:49,840 Like this is a new fundamental way of, you know, building tech 745 00:41:49,840 --> 00:41:53,720 startups is that, you know, you will become obsolete if you 746 00:41:53,720 --> 00:41:57,720 actually hold on to technology that you build for multiple 747 00:41:57,720 --> 00:41:59,920 years because all of that technology that you built two 748 00:41:59,920 --> 00:42:02,720 years back is most likely obsolete at this moment. 749 00:42:03,040 --> 00:42:06,760 And so you have to constantly like, think, you know, of this 750 00:42:06,760 --> 00:42:10,480 execution model where not only are you thinking about like new 751 00:42:10,480 --> 00:42:13,200 things to build, you're also very actively thinking about 752 00:42:13,400 --> 00:42:15,920 like, you know, things that you need to actually throw away and 753 00:42:15,920 --> 00:42:17,840 actually leverage like, you know, the innovation from the 754 00:42:17,840 --> 00:42:21,160 industry to replace that. We have another question from 755 00:42:21,160 --> 00:42:27,920 Twitter from Arsalan Khan again who says since most AI and data 756 00:42:27,920 --> 00:42:31,960 is created in English or Chinese, do you think start-ups 757 00:42:31,960 --> 00:42:35,720 should also focus on non-english or Chinese? 758 00:42:36,520 --> 00:42:39,200 And does this create a digital divide? 759 00:42:39,600 --> 00:42:43,520 1st, the content is created in many, many different languages. 760 00:42:44,240 --> 00:42:48,680 Yes, you know, I, you know, English is dominant in some 761 00:42:48,680 --> 00:42:50,880 ways, but there is, there is plenty of like, you know, 762 00:42:50,880 --> 00:42:54,000 content, plenty of systems, you know, in different languages. 763 00:42:54,000 --> 00:42:58,240 And, and I think what AI makes it possible today is it actually 764 00:42:58,240 --> 00:43:02,120 helps you build products that work like, you know, that, you 765 00:43:02,120 --> 00:43:05,120 know, that are that, that are global in nature. 766 00:43:05,320 --> 00:43:08,160 Like it's much easier today to build a product, you know, that 767 00:43:08,160 --> 00:43:12,640 you can actually bring to customers in, you know, all, all 768 00:43:12,640 --> 00:43:15,320 parts of the world. You can localize your products 769 00:43:15,320 --> 00:43:18,640 much easier with AI. You can make it work in Japanese 770 00:43:18,640 --> 00:43:23,240 and in Korean and like Hindi and all the different languages at 771 00:43:23,320 --> 00:43:26,000 at a much faster pace. Like, you know, so you know, 772 00:43:26,000 --> 00:43:31,080 this is, this is one of the, you know, a, you know, a core 773 00:43:31,080 --> 00:43:35,840 capability of AI. But then in terms of biases and 774 00:43:35,920 --> 00:43:40,440 the digital divide, it is true that the, and it's been true 775 00:43:40,440 --> 00:43:42,320 like, you know, forever, like, you know, even on the Internet, 776 00:43:42,320 --> 00:43:45,520 even Pai, when you go on Google and search for information, 777 00:43:45,880 --> 00:43:49,680 there's always that bias that creeps in because like, you 778 00:43:49,680 --> 00:43:53,400 know, English dominates as, as the source of knowledge in the 779 00:43:53,400 --> 00:43:57,200 world, right? And, and so the, that, that's 780 00:43:57,240 --> 00:43:59,240 it, that's, that's a good one. Like there's a problem I don't 781 00:43:59,240 --> 00:44:04,040 have a good answer on, like how, how to sort of resolve that as 782 00:44:04,040 --> 00:44:06,320 as as you know, AI becomes more and more capable. 783 00:44:06,320 --> 00:44:09,280 How do you make sure that it's taking everybody's point of view 784 00:44:09,280 --> 00:44:12,600 and taking, you know, all the knowledge that's out there? 785 00:44:12,600 --> 00:44:16,960 I think I think like one thing maybe I should add is that we 786 00:44:16,960 --> 00:44:20,840 have more capability today with, you know, with AI to process 787 00:44:20,840 --> 00:44:24,520 like even non digitized, you know, content in a much, much 788 00:44:24,520 --> 00:44:26,520 easier way. So hopefully, like, you know, 789 00:44:26,920 --> 00:44:29,840 there's, there's some silver lining on, on on that side that, 790 00:44:29,840 --> 00:44:35,280 you know, you can actually make use of, you know, data 791 00:44:35,280 --> 00:44:38,160 knowledge, you know, content and like different languages more 792 00:44:38,160 --> 00:44:42,680 than ever before with AI. Let's talk about start-ups now 793 00:44:42,680 --> 00:44:48,240 from the perspective of enterprise buyers, which is a a 794 00:44:48,440 --> 00:44:52,080 very important part of course of any start-ups life cycle. 795 00:44:52,080 --> 00:44:55,840 If you're, if you're an enterprise start up, do you have 796 00:44:55,840 --> 00:45:01,720 advice or a a framework that enterprise buyers can use for 797 00:45:01,720 --> 00:45:06,880 evaluating AI startups? Especially given the fact that, 798 00:45:07,000 --> 00:45:11,400 as you pointed out, every startup these days is using AI 799 00:45:11,800 --> 00:45:18,040 and the hype is so intense, very often it's, it's hard to sort 800 00:45:18,040 --> 00:45:21,160 through the claims to find out what's real and what's not. 801 00:45:21,960 --> 00:45:25,880 And I'll just mention one one thing here that I remember 802 00:45:26,160 --> 00:45:29,840 traditional software companies, ERP vendors and other enterprise 803 00:45:29,840 --> 00:45:32,320 software companies. And I have to say this 804 00:45:32,320 --> 00:45:36,600 situation, you know, 20 years ago was no different from that 805 00:45:36,600 --> 00:45:37,400 standpoint. Now. 806 00:45:37,400 --> 00:45:39,960 I mean, software companies make outlandish claims. 807 00:45:39,960 --> 00:45:43,680 And So what should IT buyers and line of business buyers do about 808 00:45:43,680 --> 00:45:45,440 it? This is one of the toughest 809 00:45:45,440 --> 00:45:49,360 problems like, you know, for, for an enterprise buyer today. 810 00:45:50,520 --> 00:45:55,560 I think the AI industry has done a bit of disservice making bold 811 00:45:55,560 --> 00:45:58,760 claims, but then not not being able to follow through. 812 00:45:59,560 --> 00:46:02,680 You know, like it's very easy to create really, really amazing 813 00:46:02,680 --> 00:46:06,640 demos and visuals, you know, for what, what AI can do for you. 814 00:46:07,520 --> 00:46:11,280 And enterprise buyers have like, you know, and they've realized 815 00:46:11,280 --> 00:46:13,920 that like, you know, as they try these products out that like, 816 00:46:13,920 --> 00:46:16,640 you know, the claims often, you know, like, you know, are are 817 00:46:16,640 --> 00:46:18,480 much larger than like, you know, the reality. 818 00:46:19,320 --> 00:46:25,400 So I think like I would say like, you know, maybe maybe let 819 00:46:25,400 --> 00:46:27,640 me take a step backwards and then first talk about like, 820 00:46:27,640 --> 00:46:29,920 well, what should we even do? Like I think like there should 821 00:46:29,920 --> 00:46:35,880 be a plan for like how you're going to roll out AI inside your 822 00:46:35,880 --> 00:46:38,840 enterprise. And I feel like, you know, 823 00:46:38,840 --> 00:46:44,520 centralizing that, you know, you know, and having a core, you 824 00:46:44,520 --> 00:46:47,520 know, AI strategy for enterprise, first of all, is the 825 00:46:47,520 --> 00:46:50,360 right way to start. Like now think about all the 826 00:46:50,360 --> 00:46:52,000 things you'd like to do this year with AI. 827 00:46:52,960 --> 00:46:55,080 What are the different areas that you would like to actually 828 00:46:55,080 --> 00:46:58,240 see AI make an impact? You can pick a few departments. 829 00:46:58,240 --> 00:47:00,840 You can say that, well, you know, for our engineering teams, 830 00:47:00,840 --> 00:47:04,160 for our customer service team, you know, like these are the top 831 00:47:04,160 --> 00:47:06,480 two or three priorities, you know, where we, you know, want 832 00:47:06,480 --> 00:47:10,200 AI to actually make an impact. So first build that road map and 833 00:47:10,200 --> 00:47:13,400 build that together in your enterprise, you know, the CIO or 834 00:47:13,400 --> 00:47:17,320 if you're just, you know, if you've chosen to have a, a head 835 00:47:17,320 --> 00:47:20,440 of AI, chief AI officer type of, you know, role in your 836 00:47:20,440 --> 00:47:22,800 enterprise, like, you know, let them, you know, give them that 837 00:47:22,800 --> 00:47:25,440 charter of like making sure they're working with all the 838 00:47:25,440 --> 00:47:27,440 different, you know, functional teams. 839 00:47:27,680 --> 00:47:31,520 And you come up with a, a desired road map for AI for you 840 00:47:31,520 --> 00:47:33,360 for this year. So you start, so you start 841 00:47:33,360 --> 00:47:38,000 there. Now in terms of like the number 842 00:47:38,000 --> 00:47:40,680 of vendors is not just like, you know, Michael, as you said, like 843 00:47:40,680 --> 00:47:44,480 it's not just the startups, it's actually every existing software 844 00:47:44,480 --> 00:47:47,480 company is also an AI company. They all have, you know, AI 845 00:47:47,480 --> 00:47:50,280 things to sell to you. And I think you to make some 846 00:47:50,280 --> 00:47:53,720 decisions there, like what's the right strategy for you? 847 00:47:54,000 --> 00:47:57,960 You know, we feel that like, like, no, like, you know, you 848 00:47:57,960 --> 00:48:01,360 have to be, you have to control like how many AI products you're 849 00:48:01,360 --> 00:48:03,640 going to actually bring in. It's very, very hard to 850 00:48:03,640 --> 00:48:05,440 evaluate. It's not easy, by the way, like, 851 00:48:05,440 --> 00:48:08,360 you know, it's just like, you know, the, the time that you 852 00:48:08,360 --> 00:48:12,400 have to spend time to evaluate every single AI product is 853 00:48:12,400 --> 00:48:14,200 enormous. Like, you know, you like, you 854 00:48:14,200 --> 00:48:17,120 know, like setting those systems up, getting them up and running 855 00:48:17,120 --> 00:48:18,960 in your environment and testing them. 856 00:48:19,320 --> 00:48:21,080 And like, you know, you sort of like companies have gone through 857 00:48:21,080 --> 00:48:24,760 this exercise where they spend 6 months, you know, and just to 858 00:48:24,760 --> 00:48:27,280 find out that, you know, that this thing didn't work. 859 00:48:27,280 --> 00:48:28,600 So like, what are the right strategies? 860 00:48:28,840 --> 00:48:32,400 So, so like 2, So, so, so my suggestions, like, you know, I 861 00:48:32,400 --> 00:48:36,400 have two pieces of advice. Like #1 like, basically work 862 00:48:36,400 --> 00:48:40,840 with fewer, you know, number. Like don't have too many POC's. 863 00:48:40,840 --> 00:48:44,680 Like, you know, start with a few so that you can do justice to 864 00:48:44,680 --> 00:48:48,480 them. And, and 2nd, instead of like 865 00:48:48,480 --> 00:48:52,480 relying on demos and, you know, getting excited by a 866 00:48:52,480 --> 00:48:55,240 presentation. Well, you can have a safer 867 00:48:55,240 --> 00:48:58,080 strategy, which is like, we'll go and what every vendor that 868 00:48:58,080 --> 00:49:01,080 you work with, like see if they have proof points like go and, 869 00:49:01,240 --> 00:49:04,400 you know, make them, you know, connect, you know, connect you 870 00:49:04,400 --> 00:49:07,200 with, you know, customers that they were able to create success 871 00:49:07,200 --> 00:49:09,400 for. So I think that is like, you 872 00:49:09,400 --> 00:49:12,400 know, it's a lot more helpful like, instead of evaluating 873 00:49:12,640 --> 00:49:15,560 like, you know, like actually had those conversations with 874 00:49:15,560 --> 00:49:18,560 your peers in the industry and see like, you know, where they 875 00:49:18,560 --> 00:49:21,200 are achieving success. Like if you tried 4 things, 876 00:49:21,200 --> 00:49:23,360 you'll see one thing that's successful, share that story 877 00:49:23,360 --> 00:49:25,760 with others and, and, and they will share with you. 878 00:49:25,760 --> 00:49:28,880 And that's probably the way to sort of scale and, and maybe 879 00:49:29,040 --> 00:49:33,320 maybe this is a plug for Glean, but the, the way we think about 880 00:49:33,320 --> 00:49:37,560 AI is that, well, like fundamentally all AI in your 881 00:49:37,560 --> 00:49:42,400 enterprise is about making, you know, working with some data 882 00:49:42,400 --> 00:49:46,440 that's in your enterprise using the reasoning and intelligence 883 00:49:46,440 --> 00:49:49,840 powers of the language models. And then after that doing some 884 00:49:49,840 --> 00:49:53,200 work, which you're again going to save, you know, in your 885 00:49:53,200 --> 00:49:55,440 enterprise systems, you know, that work is going to get 886 00:49:55,440 --> 00:49:57,280 recorded and saved in your enterprise systems. 887 00:49:57,480 --> 00:50:01,480 So, so fundamentally, when you think about all AI use cases, 888 00:50:01,480 --> 00:50:07,360 they're about working with your enterprise information and, and, 889 00:50:07,680 --> 00:50:10,400 and then you're making and applying AI on it to do some, 890 00:50:10,400 --> 00:50:12,360 you know, to, to actually make some magic happen. 891 00:50:12,640 --> 00:50:14,880 And so we chose a different strategy with Glean, which is 892 00:50:14,880 --> 00:50:17,160 that, well, why don't we actually build a system that's 893 00:50:17,160 --> 00:50:19,520 connected to all of the enterprise data, right? 894 00:50:19,520 --> 00:50:22,240 That's what Glean is. And, and now we're giving you 895 00:50:22,240 --> 00:50:26,400 this platform where you can go and build like many, many of 896 00:50:26,400 --> 00:50:28,920 these agents, many of these applications yourself. 897 00:50:30,080 --> 00:50:32,680 And, and, and that way, like, you know, with this horizontal 898 00:50:32,680 --> 00:50:35,400 strategy, you can, you know, you can do a much better job at 899 00:50:35,400 --> 00:50:38,520 security governance, you know, also like not having to buy 900 00:50:38,520 --> 00:50:41,200 many, many different products. Like it's more cost effective 901 00:50:42,120 --> 00:50:45,200 and, and and and and allows you to sort of like, you know, get 902 00:50:45,200 --> 00:50:47,200 more value through like, you know, through 1, like, you know, 903 00:50:47,200 --> 00:50:49,440 11 product. All right, we're almost out of 904 00:50:49,440 --> 00:50:51,840 time. So Arvind, I'm gonna ask you a 905 00:50:51,840 --> 00:50:54,880 bunch of questions, some from me, some from the audience 906 00:50:54,880 --> 00:50:59,840 listening and I'll ask you to respond very quickly from an 907 00:50:59,840 --> 00:51:04,280 enterprise perspective, the build versus buy decision, how 908 00:51:04,280 --> 00:51:06,200 should enterprises make that choice? 909 00:51:06,320 --> 00:51:09,240 Very quickly please. It's a build plus buy. 910 00:51:09,880 --> 00:51:14,200 You have to make sure that you get as much turnkey technology 911 00:51:14,200 --> 00:51:17,320 as you can, but you have to also remember that it's not going to 912 00:51:17,320 --> 00:51:19,280 be enough like, you know, to add true value. 913 00:51:19,440 --> 00:51:21,720 You will have to, you know, you know, build on top of those 914 00:51:21,720 --> 00:51:26,040 systems. From Twitter, a big hurdle for 915 00:51:26,160 --> 00:51:30,920 AI and LLM platforms is that they are not integrated with 916 00:51:30,960 --> 00:51:32,920 work flows. How do you see the market 917 00:51:32,960 --> 00:51:34,760 evolving? That's exactly right. 918 00:51:34,760 --> 00:51:39,000 And I think the you you have to, you know, build that layer. 919 00:51:39,600 --> 00:51:43,080 One of the key techniques for that has been rag like so you 920 00:51:43,080 --> 00:51:46,320 take all of enterprise data and knowledge systems for flows and 921 00:51:46,320 --> 00:51:49,600 you actually build the, you know this, you know, a retrieval 922 00:51:50,080 --> 00:51:53,440 system and an action system on top of all of your enterprise 923 00:51:53,440 --> 00:51:56,000 data and systems. And then you connect with AI 924 00:51:56,000 --> 00:51:57,720 like you know, that's exactly what we, you know, the problem 925 00:51:57,720 --> 00:52:02,760 that we solve with you. Another one from Twitter who the 926 00:52:02,800 --> 00:52:08,200 CTOCFOCOO chief digital officer? Who creates AI strategy? 927 00:52:08,200 --> 00:52:12,720 And should the chief AI officer report directly to the CICEO? 928 00:52:13,280 --> 00:52:15,720 That's a good idea. Like so if you, if you actually 929 00:52:15,880 --> 00:52:22,600 have a chief AI officer, having them report to the the CIO or 930 00:52:22,600 --> 00:52:26,080 the CEO could be a good strategy because this is a company wide 931 00:52:26,080 --> 00:52:27,880 effort. So like the, if you don't 932 00:52:27,880 --> 00:52:29,960 associate it with a function that's actually better in my 933 00:52:29,960 --> 00:52:32,920 opinion. And but otherwise, like, you 934 00:52:32,920 --> 00:52:35,880 know, I think like, like, you know, it's not don't be rigid. 935 00:52:35,880 --> 00:52:40,120 Like, you know, like whoever is motivated, who's excited and 936 00:52:40,120 --> 00:52:43,240 like, you know, like and has the capability, let them drive their 937 00:52:43,280 --> 00:52:45,680 efforts initially And like if you can always figure out how to 938 00:52:45,680 --> 00:52:48,000 reorganize later, like, you know, in a more scalable way. 939 00:52:48,480 --> 00:52:51,160 Here we have another question from Twitter, and this is about 940 00:52:51,160 --> 00:52:54,960 pain. If AI startups are getting more 941 00:52:54,960 --> 00:52:59,520 lean, are older AI startups at a disadvantage? 942 00:52:59,920 --> 00:53:04,800 And as these older startups become more lean because more 943 00:53:04,800 --> 00:53:08,480 efficient using AI, what is the implication for employment and 944 00:53:08,480 --> 00:53:11,080 their workforce? So on, on the 1st question, 945 00:53:11,080 --> 00:53:15,720 there is indeed a like a first mover disadvantage in AI because 946 00:53:15,720 --> 00:53:19,080 when technology changes so fast and you, you have some of that 947 00:53:19,080 --> 00:53:22,440 legacy code base in your systems, like it makes it like, 948 00:53:22,640 --> 00:53:23,920 makes you a little bit less agile. 949 00:53:23,920 --> 00:53:26,560 So well, that's, that's part of like, you know how it always is. 950 00:53:26,560 --> 00:53:28,840 Like, you know, as a new startup, you always have the 951 00:53:28,840 --> 00:53:31,160 agility advantage. And as a start, that's a little 952 00:53:31,160 --> 00:53:32,720 bit older. Well, like, you know, work hard, 953 00:53:32,720 --> 00:53:35,760 like work hard on modernizing, you know, your systems, your 954 00:53:35,760 --> 00:53:38,920 stacks so that you can, so you don't fall behind for those 955 00:53:38,920 --> 00:53:41,600 reasons. And then in terms of employment, 956 00:53:41,760 --> 00:53:44,960 I think the like, I'll, I'll just say one thing. 957 00:53:45,160 --> 00:53:49,440 I've not seen people losing employment because of AI. 958 00:53:49,760 --> 00:53:52,480 Yeah, like, you know, we work with so many large enterprises. 959 00:53:52,640 --> 00:53:55,560 They're bringing so much automation and efficiency for 960 00:53:55,560 --> 00:53:57,240 them. But like, you know, every, like, 961 00:53:57,240 --> 00:53:59,440 you know, every enterprise that we work with, you know, they 962 00:53:59,440 --> 00:54:02,840 actually are concerned with both bottom line and top line. 963 00:54:02,960 --> 00:54:05,800 And so when they get some bins, nobody's, you know, giving up 964 00:54:05,800 --> 00:54:07,280 their team members. Like, you know, they're not 965 00:54:07,280 --> 00:54:10,640 actually, you know, they're just actually thinking about like, 966 00:54:10,640 --> 00:54:13,080 well, I can do more. And, and so companies are 967 00:54:13,080 --> 00:54:15,240 actually building products at a faster pace. 968 00:54:15,240 --> 00:54:16,760 They're getting more things done. 969 00:54:17,640 --> 00:54:19,720 So, so I'm not, I'm not super worried like, you know, I think 970 00:54:19,720 --> 00:54:22,120 I guess like, like I will come back to the point that well, 971 00:54:22,120 --> 00:54:24,040 like just stay relevant. You know, that's that's the 972 00:54:24,040 --> 00:54:25,760 thing that matters for you as an individual. 973 00:54:26,520 --> 00:54:29,680 If you if you learn how to use AI you know you will have no 974 00:54:29,680 --> 00:54:34,320 problems you know in the future. You're advocating maintaining 975 00:54:34,320 --> 00:54:37,080 intellectual curiosity. I'm not trying to put words in 976 00:54:37,080 --> 00:54:42,840 your mouth because as AI drives efficiency gains, if you're 977 00:54:42,840 --> 00:54:48,520 intellectually current on what's going on, you can adapt and work 978 00:54:48,520 --> 00:54:51,200 around that the changes that are taking place. 979 00:54:51,600 --> 00:54:53,160 Yeah, and companies are really hungry. 980 00:54:53,160 --> 00:54:57,160 They're really hungry for AI experts, for AI talents or like, 981 00:54:57,160 --> 00:54:59,080 so become one. I think it's going to be good 982 00:54:59,080 --> 00:55:00,840 for you. What advice do you have for 983 00:55:00,840 --> 00:55:03,760 entrepreneurs considering starting an AI startup? 984 00:55:04,080 --> 00:55:05,920 Number one conviction like you know, we know. 985 00:55:05,920 --> 00:55:10,600 Stay, stick with your idea, don't give up and, and #2 well, 986 00:55:10,600 --> 00:55:12,920 like, you know, if you're, if you want to succeed, like, you 987 00:55:12,920 --> 00:55:15,360 know, be ready to, to work very, very hard. 988 00:55:16,280 --> 00:55:20,480 And, and, and then, then the last thing I would say is that, 989 00:55:20,680 --> 00:55:23,400 you know, a company is all about people that you, you know, build 990 00:55:23,400 --> 00:55:26,480 it with. So like, you know, focus on like 991 00:55:26,480 --> 00:55:30,560 having, you know, having a great following team and, and, and 992 00:55:30,560 --> 00:55:33,600 spend a lot of time, you know, trying to get the right people 993 00:55:33,600 --> 00:55:35,520 that employees, you know, in your organization. 994 00:55:35,760 --> 00:55:38,920 And like you know, when you get the right people, you know they 995 00:55:38,920 --> 00:55:40,560 will do the you know they will build great products. 996 00:55:40,560 --> 00:55:43,600 They will make you succeed. Final thoughts or advice for 997 00:55:43,600 --> 00:55:46,720 CIOs who are making AI investment decisions very 998 00:55:46,720 --> 00:55:52,280 quickly, please. Focus on security, focus on the 999 00:55:53,760 --> 00:55:57,920 like, you know, centralizing, you know, like your AI software 1000 00:55:57,920 --> 00:56:03,640 stack and, and, and, and I think the, the, the last thing is, you 1001 00:56:03,640 --> 00:56:08,400 know, that I would say is that like we don't try to sort of get 1002 00:56:08,400 --> 00:56:10,400 small wins. You know, this is my advice. 1003 00:56:10,400 --> 00:56:12,800 Like, you know, don't like, you know, don't like, you know, like 1004 00:56:12,800 --> 00:56:16,160 create projects where you know, if it doesn't deliver 50% 1005 00:56:16,160 --> 00:56:18,560 success, it's a failure. Like, you know, like, you know, 1006 00:56:18,560 --> 00:56:21,280 force every, you know, team, every function in your team, 1007 00:56:21,320 --> 00:56:24,480 like in your organization to, to pick, you know, one or two wins 1008 00:56:24,480 --> 00:56:27,600 for AI, like, you know, every quarter and see how that goes. 1009 00:56:28,080 --> 00:56:30,640 All right. And with that, we are out of 1010 00:56:30,640 --> 00:56:34,280 time. A huge thank you to Arvind Jane 1011 00:56:34,280 --> 00:56:37,120 from Glean. Arvind, thank you so much for 1012 00:56:37,120 --> 00:56:38,640 taking your time to be with us today. 1013 00:56:39,280 --> 00:56:40,840 Thank you so much, it was a lot of fun. 1014 00:56:41,320 --> 00:56:44,360 And thanks to everybody who watched and especially you folks 1015 00:56:44,360 --> 00:56:47,880 who asked amazing questions. You guys are truly awesome. 1016 00:56:48,120 --> 00:56:51,520 Before you go, subscribe to our newsletter. 1017 00:56:51,520 --> 00:56:54,640 Join our community. Go to cxotalk.com. 1018 00:56:54,880 --> 00:56:59,200 Subscribe to the newsletter, check it out and we'll see you 1019 00:56:59,200 --> 00:57:02,280 again next time everybody. We have awesome shows coming. 1020 00:57:02,280 --> 00:57:04,240 U talking about AI? See you later.