1 00:00:00,120 --> 00:00:04,280 I'm fresh off the Mongo DB dot local event. 2 00:00:04,360 --> 00:00:07,160 They had a ton of developers, partners, startup founders 3 00:00:07,160 --> 00:00:10,400 hanging around and showing how Mongo is trying to respond to 4 00:00:10,400 --> 00:00:13,080 everything that's happening in artificial intelligence. 5 00:00:13,120 --> 00:00:16,200 They are the sponsor of this podcast and they had me on site 6 00:00:16,200 --> 00:00:19,560 at Mongo DB dot local to do a bunch of these interviews, which 7 00:00:19,880 --> 00:00:22,440 I really enjoyed, so we decided to put them in our feed. 8 00:00:22,440 --> 00:00:27,280 I had a great conversation with CJ Desai, the CEO of Mongo, who 9 00:00:27,280 --> 00:00:29,960 dropped by my little studio at the conference. 10 00:00:30,040 --> 00:00:32,840 We have another interview with the head of AI from Ripling 11 00:00:32,840 --> 00:00:35,520 coming after that. This is the Newcomer podcast. 12 00:00:43,280 --> 00:00:46,080 Hi, I'm Eric Newcomer, author of Newcomer. 13 00:00:46,320 --> 00:00:52,000 We're here at Mongo DB's local event, the amazing drop in when 14 00:00:52,000 --> 00:00:56,960 you get the CEO, what, 65 days into the job to show up on 15 00:00:56,960 --> 00:00:59,720 stage. What I mean, this is, this is 16 00:00:59,720 --> 00:01:02,080 your big event. What was What's the message that 17 00:01:02,080 --> 00:01:04,959 you really wanted to carry to the attendees here? 18 00:01:05,440 --> 00:01:09,320 You know, I would say, Eric, first, thank you for having me 19 00:01:09,400 --> 00:01:13,400 and thank you to you for letting me crash the party of. 20 00:01:13,400 --> 00:01:15,800 Course I love it. So we. 21 00:01:15,800 --> 00:01:18,280 Just we just literally we just talked to a three person 22 00:01:18,800 --> 00:01:21,920 company. So now we're going much larger. 23 00:01:21,920 --> 00:01:23,280 Yeah. Sounds good, Eric. 24 00:01:23,280 --> 00:01:29,960 So 65 days in and one of the things that our previous CEO, 25 00:01:30,000 --> 00:01:35,040 Dev and the entire team, we realized, so first of all, from 26 00:01:35,960 --> 00:01:40,280 just being in Silicon Valley for the long time, San Francisco 27 00:01:40,280 --> 00:01:42,880 feels like it's back. San Francisco is back. 28 00:01:43,080 --> 00:01:46,640 During the pandemic, people went a little dark on San Francisco. 29 00:01:46,960 --> 00:01:52,720 But with this AI platform shift, San Francisco is back and as you 30 00:01:52,720 --> 00:01:55,360 know, we had a New York headquarter, New York founded 31 00:01:55,360 --> 00:01:57,640 company. Listen, I'm I'm a New York based 32 00:01:57,640 --> 00:01:59,640 person who also believes in San Francisco. 33 00:01:59,640 --> 00:02:03,120 So we have a sort of shared spirit on that realizing that AI 34 00:02:03,120 --> 00:02:05,880 boom is here, we are also in New York. 35 00:02:05,880 --> 00:02:10,240 So yeah, I feel it. So lot of you know, so San 36 00:02:10,240 --> 00:02:14,120 Francisco is back. Mongo DB about 10 years ago did 37 00:02:14,120 --> 00:02:15,840 a really nice job at San Francisco. 38 00:02:16,160 --> 00:02:20,560 They were in front of software developers, builders saying 39 00:02:20,560 --> 00:02:22,240 build on Mongo DB. Here is why. 40 00:02:22,240 --> 00:02:26,160 Speed, agility, scale out many, many advantages. 41 00:02:26,600 --> 00:02:32,160 And then the team realized, Dave and the team that we kind of as 42 00:02:32,160 --> 00:02:35,200 the company succeeded, we took our eye off the ball for a lack 43 00:02:35,200 --> 00:02:40,400 of better term. And so after four years, we 44 00:02:40,400 --> 00:02:44,800 decided to reintroduce Mongo DB today in San Francisco. 45 00:02:44,960 --> 00:02:49,720 So today's an auspicious day. Thank you for coming and we 46 00:02:49,720 --> 00:02:53,280 wanted to launch and tell people that we are here, we are Mongo 47 00:02:53,280 --> 00:02:54,840 DB. It's a great data platform you 48 00:02:54,840 --> 00:02:57,880 can build on no matter what kind of applications you're building 49 00:02:57,880 --> 00:03:01,680 on, digital native, AI native or you have AI plus plus, whatever 50 00:03:01,680 --> 00:03:04,040 the case might be, it is a great database. 51 00:03:04,040 --> 00:03:09,160 And you know, one of my biggest profound realization was that 52 00:03:09,160 --> 00:03:13,040 when I, when I was doing my own diligence to join Mongo DB, even 53 00:03:13,560 --> 00:03:16,840 though the founders didn't create with AI in mind, 54 00:03:17,320 --> 00:03:20,960 unstructured data, flexible schema, you know, Symantec 55 00:03:22,600 --> 00:03:25,360 Retrieval and all these things that are part of now the 56 00:03:25,360 --> 00:03:29,960 platform, it is like the platform for AI applications. 57 00:03:29,960 --> 00:03:32,000 And so we wanted to tell that to everyone. 58 00:03:32,000 --> 00:03:36,040 And we got a great customer today to validate a great 59 00:03:36,040 --> 00:03:38,920 founder who said he has built three companies on Mocodb. 60 00:03:39,560 --> 00:03:41,120 You're talking about Mike Krieger. 61 00:03:41,440 --> 00:03:46,440 Yeah, you had the nice self video Instagram founder. 62 00:03:46,840 --> 00:03:49,800 I created artifacts, which you talked about briefly, which is 63 00:03:49,800 --> 00:03:51,960 an interesting for me in the news world. 64 00:03:51,960 --> 00:03:54,400 And then obviously he's at Anthropic. 65 00:03:54,400 --> 00:03:56,960 Yeah, that was an exciting endorsement. 66 00:03:58,120 --> 00:04:01,560 What what is your read on? You know, I don't know, AI hype 67 00:04:01,560 --> 00:04:03,800 in 2026. On some level, you must think 68 00:04:03,800 --> 00:04:07,280 this is obviously a phenomenon to say you're doing this big San 69 00:04:07,280 --> 00:04:10,320 Francisco push. Dave, you know, I think I talked 70 00:04:10,320 --> 00:04:12,480 to him last year. You know, he has a dose of 71 00:04:12,480 --> 00:04:15,000 realism to it all. What's your what's your personal 72 00:04:15,000 --> 00:04:17,279 perspective on where we are in this hype cycle? 73 00:04:17,279 --> 00:04:21,560 He's smart and realist and I'm an optimist, so that's how I. 74 00:04:21,600 --> 00:04:24,360 Brought in the guy like, Oh no, this thing is going strong and 75 00:04:24,360 --> 00:04:25,800 we're we're going to get in front of it. 76 00:04:25,960 --> 00:04:28,960 Yeah. I mean, I, I would say since 77 00:04:28,960 --> 00:04:32,840 2022, Christmas ish or October ish, when you look at the 78 00:04:32,840 --> 00:04:37,000 evolution of AI and you look at some of the killer AI companies 79 00:04:37,000 --> 00:04:42,640 where they have scale business in a significant way has been of 80 00:04:42,640 --> 00:04:44,880 course open AI. Everybody understands that that 81 00:04:44,880 --> 00:04:47,520 is a killer app from my perspective. 82 00:04:48,280 --> 00:04:52,440 But you also look at Grog and how fast they have grown X AI 83 00:04:52,880 --> 00:04:58,760 and then Entropic, the entire team at Entropic Lab, that is a 84 00:04:58,760 --> 00:05:04,200 truly killer AI app for coders. I mean, people load that 85 00:05:04,200 --> 00:05:07,560 platform and they sometimes use it with cursor this that. 86 00:05:07,840 --> 00:05:11,240 So you're seeing that these companies, if you look at like 87 00:05:12,240 --> 00:05:16,520 Internet age in 90s, you look at mobile age, they did not scale 88 00:05:16,520 --> 00:05:19,600 this fast. How how fast this companies have 89 00:05:19,600 --> 00:05:22,840 scaled. That's like pretty amazing from 90 00:05:22,840 --> 00:05:26,120 my standpoint. And in 2025, I think the killer 91 00:05:26,120 --> 00:05:29,960 apps were the coding tools. There was a killer app besides 92 00:05:29,960 --> 00:05:33,560 chat GPP of course. And now as we go into 2026, 93 00:05:34,200 --> 00:05:35,680 which ones are going to take off? 94 00:05:36,640 --> 00:05:39,880 I'm optimistic that some very specialized vertical apps 95 00:05:40,160 --> 00:05:44,760 related to maybe a healthcare or insurance insurance assist or 96 00:05:44,760 --> 00:05:48,160 something may take off, but you will see some take. 97 00:05:48,160 --> 00:05:50,600 Off close to, you know, a bridge and open evidence. 98 00:05:50,600 --> 00:05:52,920 Those are super interesting applications. 99 00:05:52,920 --> 00:05:55,400 And health, Yeah. You interviewed Constantine and 100 00:05:55,400 --> 00:05:56,480 Sequoia. Obviously, they're big. 101 00:05:56,800 --> 00:05:58,440 They're bullish on Harvey, you know. 102 00:05:58,480 --> 00:06:00,480 Yeah. So a lot of these vertical 103 00:06:00,480 --> 00:06:04,840 applications are exciting. And Harvey, you know, I met them 104 00:06:04,840 --> 00:06:10,160 a couple of years ago and they were still, you know, truly with 105 00:06:10,160 --> 00:06:12,840 the legal firms they were using them. 106 00:06:13,240 --> 00:06:16,840 And then now you have the in house counsels use them as well. 107 00:06:17,120 --> 00:06:20,560 And so they have gotten a perfect product market fit and 108 00:06:20,560 --> 00:06:22,280 they're expanding now their use cases. 109 00:06:22,280 --> 00:06:24,320 So that's just it. But that's very specialized, 110 00:06:24,320 --> 00:06:25,920 right? It's not just generic. 111 00:06:25,920 --> 00:06:30,320 I come in and help me review this contract from a legality 112 00:06:30,320 --> 00:06:33,280 perspective and so on. This is a drop by, so I don't 113 00:06:33,280 --> 00:06:34,400 want to take too much of your time. 114 00:06:34,400 --> 00:06:37,800 This will be the last question. How do you think about, you 115 00:06:37,800 --> 00:06:40,280 know, there's so many models available. 116 00:06:40,280 --> 00:06:45,160 How do you think about like, oh, where to provide your own versus 117 00:06:45,160 --> 00:06:48,040 to just sort of say you're obviously going to bring models 118 00:06:48,040 --> 00:06:50,720 from other places? We provide sort of the database 119 00:06:50,720 --> 00:06:54,280 layer. For databases, we want to be 120 00:06:54,280 --> 00:06:57,200 model agnostic, Yeah, we want to provide best embeddings. 121 00:06:57,200 --> 00:07:00,560 So your retrieval quality is high, accuracy is high, but we 122 00:07:00,560 --> 00:07:04,440 want to be model agnostic. And when I speak to customers, 123 00:07:04,680 --> 00:07:07,720 including AI companies, like everybody who originally was 124 00:07:07,720 --> 00:07:12,280 telling me open AI, then they shifted to cloud, I think the 125 00:07:12,280 --> 00:07:15,720 innovation cycles are very fast. And then you look at some of the 126 00:07:15,720 --> 00:07:18,960 firms in China and you look at DeepSeek and others that 127 00:07:18,960 --> 00:07:20,480 innovation. Cycles come from anywhere. 128 00:07:20,560 --> 00:07:24,400 Yeah, that could come from and the product cycles are shorter, 129 00:07:24,640 --> 00:07:26,320 right. Product cycles are shorter. 130 00:07:26,600 --> 00:07:32,400 So multi model, not multi modal, but multi model is will be a way 131 00:07:32,400 --> 00:07:37,440 to go and it gives freedom to people to use whichever model is 132 00:07:37,440 --> 00:07:39,120 best. Gemini, maybe today, better 133 00:07:39,120 --> 00:07:42,000 tomorrow, maybe an topic again. And I think that's how it's 134 00:07:42,000 --> 00:07:44,120 going to work. So we will be agnostic 135 00:07:44,920 --> 00:07:48,400 regardless of who you use. Our goal would be always to give 136 00:07:48,880 --> 00:07:54,080 best semantic retrieval capabilities and completely 137 00:07:54,080 --> 00:07:57,960 provide a scalable data plan. Well, I love, you know, 138 00:07:58,160 --> 00:08:01,320 newcomer, we use green so it was easy to share the stage, but 139 00:08:01,320 --> 00:08:03,880 honoured that we get to share branding with Mongo. 140 00:08:03,880 --> 00:08:06,320 Thanks so much for having me here and thanks for dropping by. 141 00:08:06,360 --> 00:08:08,440 Thank you, Our stage. We appreciate it and we'll see 142 00:08:08,440 --> 00:08:09,800 you in Brooklyn soon. Sounds good. 143 00:08:10,080 --> 00:08:12,920 Alright, thank you. Thanks again to Mongo DB for 144 00:08:12,920 --> 00:08:15,600 sponsoring this episode. I feel like I've talked to both 145 00:08:15,600 --> 00:08:17,840 CEOs, the old and the new in the last six months. 146 00:08:17,840 --> 00:08:20,880 So that's been a lot of fun. And now excited to get into it 147 00:08:20,880 --> 00:08:23,000 with Anchor Bot, the head of AI at Ripling. 148 00:08:23,040 --> 00:08:25,520 I know he's big boss Parker Conrad pretty well. 149 00:08:25,520 --> 00:08:27,240 He's been on this podcast before. 150 00:08:27,240 --> 00:08:29,880 Parker has been somewhat slow to embrace AI. 151 00:08:29,880 --> 00:08:32,720 So it's fun to talk to his head of AI about, you know, the 152 00:08:32,720 --> 00:08:37,200 company finding AI Jesus, believing in AIA little bit and 153 00:08:37,400 --> 00:08:40,360 how they're making that cultural and technological 154 00:08:40,360 --> 00:08:42,559 transformation. Give it a listen. 155 00:08:42,559 --> 00:08:46,440 Thrilled to have anchor bot, the head of AI at Rippling here. 156 00:08:46,440 --> 00:08:49,520 I'm just saying, before we got on, I go way back with Parker 157 00:08:49,520 --> 00:08:52,800 Conrad, your Ceoi wrote about Zenefits back in the day. 158 00:08:52,800 --> 00:08:55,480 I I feel like I was, you know, as far as reporters can be 159 00:08:55,480 --> 00:08:57,160 bullish. I was pretty bullish on his 160 00:08:57,480 --> 00:09:00,000 comeback in Rippling and have been following your work. 161 00:09:01,280 --> 00:09:04,200 What's it like to be the head of AI for? 162 00:09:04,200 --> 00:09:08,520 I feel like for a Parker is a technologist for the guy, but he 163 00:09:08,520 --> 00:09:10,720 has not been wrapping himself in AI. 164 00:09:10,720 --> 00:09:14,840 Like how's that How how is it to be the head of AI and CEO sort 165 00:09:14,840 --> 00:09:19,240 of coming to the the AI hype? I think it's been exciting 166 00:09:19,240 --> 00:09:22,600 specifically from the lens of there is so much happening in AI 167 00:09:22,680 --> 00:09:26,760 everyday. So automatically everybody's 168 00:09:26,760 --> 00:09:30,240 curiosity on what's the new model release, what's the new 169 00:09:30,240 --> 00:09:32,920 capability release? And like like these days, 170 00:09:32,920 --> 00:09:36,320 everybody's hyped up on cloud code as an example, what 171 00:09:36,320 --> 00:09:39,840 Entropic is doing. So that automatically starts to 172 00:09:39,840 --> 00:09:42,680 create this pull internally around. 173 00:09:42,680 --> 00:09:46,040 What does this mean for Rippling from a product point of view? 174 00:09:46,440 --> 00:09:49,640 What does it mean for Rippling in terms of day-to-day business 175 00:09:50,000 --> 00:09:51,960 perspective? And what does it mean in terms 176 00:09:51,960 --> 00:09:55,000 of customers expectation on Rippling it, right? 177 00:09:55,520 --> 00:09:57,600 You get to you get to be the guy who knows what's happening 178 00:09:57,600 --> 00:09:59,520 today. I say, oh, how do we respond to 179 00:09:59,520 --> 00:10:02,160 this, this, that and the other we're obviously, you know, at 180 00:10:02,160 --> 00:10:07,600 Mongo DB dot local, you know, obviously sprawling ambition 181 00:10:07,600 --> 00:10:10,000 with Rippling, it's like, oh, we want to stack startup on 182 00:10:10,000 --> 00:10:13,760 startup. So I can imagine strong data 183 00:10:13,760 --> 00:10:17,280 organization capability is part of the business success. 184 00:10:17,280 --> 00:10:19,920 But explain Ripling's relationship with Mongo. 185 00:10:19,960 --> 00:10:21,480 Yeah. So we are being a long time 186 00:10:21,480 --> 00:10:24,080 customer of Mongo from the beginning. 187 00:10:24,080 --> 00:10:27,800 I think as you rightly pointed out, Parker's thesis was 188 00:10:28,680 --> 00:10:32,120 compound startup like product over product. 189 00:10:32,560 --> 00:10:36,320 And the core, heart of it is the employee graph. 190 00:10:36,800 --> 00:10:41,160 And being even though yes people know us that we can run payroll 191 00:10:41,240 --> 00:10:44,120 for you. Core of it is the employee graph 192 00:10:44,120 --> 00:10:48,560 which captures not just your pay related information or benefits 193 00:10:48,560 --> 00:10:51,920 related information. We also capture your ID and 194 00:10:51,920 --> 00:10:54,760 identity related information, your device related information, 195 00:10:54,760 --> 00:10:58,480 because at the end we also have a product portfolio of ID. 196 00:10:59,160 --> 00:11:02,160 Similarly, we also have a product portfolio around 197 00:11:03,040 --> 00:11:06,240 finance, your corporate card, your spend, your travel expense. 198 00:11:06,240 --> 00:11:08,120 We just launched a travel expense product. 199 00:11:08,560 --> 00:11:12,360 So that's certainly the amount of data you have about employees 200 00:11:12,840 --> 00:11:15,720 gets connected in so many different ways. 201 00:11:16,080 --> 00:11:20,320 And having a partnership with Mongo allows us to continue 202 00:11:20,920 --> 00:11:24,920 keeping that graph sanity intact in terms of the relationships 203 00:11:24,920 --> 00:11:30,160 you have in this business data. Rigorous are you guys about 204 00:11:30,160 --> 00:11:33,200 making sure that everything built at Ripling uses the same 205 00:11:33,280 --> 00:11:35,280 tech stack? I mean, if you're acquiring 206 00:11:35,280 --> 00:11:37,720 startups and you're letting people sort of do their thing, 207 00:11:38,080 --> 00:11:42,440 how do you balance the trade off of sort of a consistent stack 208 00:11:42,440 --> 00:11:44,680 versus letting different people build what they want? 209 00:11:44,840 --> 00:11:48,600 Oh, that's a great question. And speed versus consistency is 210 00:11:48,600 --> 00:11:50,560 a constant dialogue we have internally. 211 00:11:51,000 --> 00:11:56,040 I think there is a value in terms of letting people innovate 212 00:11:56,040 --> 00:12:00,680 on the side, but it's a little bit of a trade off because as 213 00:12:00,680 --> 00:12:04,640 soon as they are reusing the common capabilities we already 214 00:12:04,640 --> 00:12:09,240 have in our platform, there is a dimensional speed which comes 215 00:12:09,240 --> 00:12:12,680 with it. And certainly for example, the 216 00:12:12,720 --> 00:12:15,360 employee graph comes with the workflow, it comes with the 217 00:12:15,360 --> 00:12:19,240 notification, it comes with an ability to you know, track the 218 00:12:19,240 --> 00:12:22,640 changes that's happening now. If I am building now our travel 219 00:12:22,640 --> 00:12:25,480 expense product that are workflows for me. 220 00:12:26,160 --> 00:12:30,080 Of course there is a dimension where I need to go into 221 00:12:30,280 --> 00:12:33,800 integrating into flight booking, hotel booking, which is very 222 00:12:33,800 --> 00:12:39,160 unique to my products niche. And there I'm free to choose 223 00:12:39,360 --> 00:12:41,440 what I choose to do that and run faster. 224 00:12:42,040 --> 00:12:45,600 But I think we've been able to guide the different product 225 00:12:45,600 --> 00:12:49,160 teams even though running independently on the core value 226 00:12:49,760 --> 00:12:53,160 of building on that common employee graph and common 227 00:12:53,160 --> 00:12:55,960 capabilities. What has been the value of Mongo 228 00:12:55,960 --> 00:12:57,320 like? Would would Rippling be a 229 00:12:57,320 --> 00:13:00,200 different company who is built on Postgres? 230 00:13:00,640 --> 00:13:04,720 I think, I think it would, no, Rippling would not be a 231 00:13:04,720 --> 00:13:08,400 different company because of Parker's overall hypothesis of a 232 00:13:08,400 --> 00:13:12,080 compound startup and really building it as a connected 233 00:13:12,120 --> 00:13:15,200 ecosystem of products running on a common foundation. 234 00:13:15,640 --> 00:13:19,200 I think it would have been just a little bit harder, I think and 235 00:13:19,240 --> 00:13:23,080 having that partnership with Mongo from day one allowed us to 236 00:13:23,160 --> 00:13:28,640 build this employee graph, build a capability set so that the 237 00:13:28,640 --> 00:13:31,960 next product or the next product we were creating was a lot more 238 00:13:31,960 --> 00:13:33,840 easy. And this is not just about 239 00:13:33,960 --> 00:13:37,960 transactional data which we are maintaining and managing. 240 00:13:38,120 --> 00:13:39,840 It's also about analytics and reporting. 241 00:13:40,080 --> 00:13:43,400 Being able to take that information and have the ability 242 00:13:43,400 --> 00:13:47,080 to report across a cross section of your product portfolio or 243 00:13:47,240 --> 00:13:52,440 your business data, not just from HR or IT or finance allowed 244 00:13:52,720 --> 00:13:56,600 has opened up those possibilities which having a 245 00:13:56,600 --> 00:14:02,560 common data store, having this doc Mongo capabilities opened up 246 00:14:02,560 --> 00:14:05,640 those possibilities for us, which we would have solved 247 00:14:05,640 --> 00:14:08,040 eventually. It would just have taken a lot 248 00:14:08,040 --> 00:14:10,600 more effort. So you're the head of AI. 249 00:14:10,600 --> 00:14:14,040 We touched on some of that role means translating what's 250 00:14:14,040 --> 00:14:17,520 happening in the broader San Francisco ecosystem to to 251 00:14:17,520 --> 00:14:19,200 Rippling. But in terms of, you know, 252 00:14:19,200 --> 00:14:23,160 products within Rippling, what what are the AI products? 253 00:14:23,160 --> 00:14:25,480 Do you have an agent yet? What like what's that look like 254 00:14:25,480 --> 00:14:27,160 inside of Rippling? Great question. 255 00:14:27,360 --> 00:14:30,120 I think originally Rippling also started. 256 00:14:30,240 --> 00:14:34,840 We have almost 70 plus products in the portfolio and originally 257 00:14:34,840 --> 00:14:39,560 each product started innovating around how they can embed AI as 258 00:14:39,560 --> 00:14:41,040 they're building that product, right. 259 00:14:41,040 --> 00:14:43,680 It could be within our recruiting product portfolio. 260 00:14:43,680 --> 00:14:46,040 When you're doing interviews, you are meeting candidates 261 00:14:46,040 --> 00:14:50,560 summarizing those information. It could be within our IIT 262 00:14:50,560 --> 00:14:53,920 product portfolio where you are issuing devices, you are 263 00:14:54,040 --> 00:14:58,120 tracking compliance and security around those things and having a 264 00:14:58,120 --> 00:15:01,560 summary of your policies. So people started embedding AI 265 00:15:01,840 --> 00:15:03,920 as product capabilities very early. 266 00:15:04,040 --> 00:15:07,320 Inadvertently, I think the change happened. 267 00:15:07,600 --> 00:15:11,480 The point in time we actively started investing on for 268 00:15:11,480 --> 00:15:14,360 building AI products, you have to use AI everyday. 269 00:15:14,880 --> 00:15:18,720 Without that, I genuinely believe it's very hard for 270 00:15:18,720 --> 00:15:20,880 people to reimagine what it means for their product. 271 00:15:21,240 --> 00:15:27,240 So 10 months ago, we Albert, our CTO really put a charter out and 272 00:15:27,240 --> 00:15:31,320 that's where I came in is driving the AI transformation of 273 00:15:31,320 --> 00:15:37,040 Rippling internally that we use AI everyday, we embrace latest 274 00:15:37,040 --> 00:15:40,120 and greatest of AI tools every day in our workflow whether it 275 00:15:40,120 --> 00:15:44,600 is product design, engineering, legal, finance, sales, marketing 276 00:15:45,240 --> 00:15:47,480 and that. What is that cursor? 277 00:15:47,480 --> 00:15:49,800 Harvey Sierra? What do? 278 00:15:49,960 --> 00:15:53,160 You use the tech stack is full of all the tools you can think 279 00:15:53,160 --> 00:15:56,360 of, right? So of course Gemini chat GBD 280 00:15:56,360 --> 00:16:01,840 cursor cloud codecs, so does. That engineer sort of pick their 281 00:16:01,840 --> 00:16:04,600 preferred. I think so because Rippling 282 00:16:04,600 --> 00:16:07,760 deals with a lot of enterprise customer data. 283 00:16:08,000 --> 00:16:14,280 We do have AI pilot process and I have built up a checklist with 284 00:16:14,280 --> 00:16:18,960 our security and legal so that we can quickly assess a new AI 285 00:16:18,960 --> 00:16:20,320 solution or somebody will request. 286 00:16:20,320 --> 00:16:23,720 So I get request every week. OK, I want to try granola. 287 00:16:23,720 --> 00:16:27,320 I want to try Bisper flow. Typically what we do is we run 288 00:16:27,320 --> 00:16:29,120 through it through an AI pilot process. 289 00:16:29,120 --> 00:16:31,080 Do you know do you use factory? I'm going to be talking to the 290 00:16:31,080 --> 00:16:34,400 factory. It's on my list of an ask from 291 00:16:34,400 --> 00:16:35,840 somebody. Maybe it'll get accelerated 292 00:16:35,840 --> 00:16:37,560 after yes. That could be. 293 00:16:37,560 --> 00:16:41,360 I'm very keen to learn more, which keeps my job a little 294 00:16:41,600 --> 00:16:44,640 curious because a lot of times I can't stay on top of things. 295 00:16:44,640 --> 00:16:48,280 So then the rest of the company is constantly keeping top of 296 00:16:48,280 --> 00:16:50,760 things and asking me can we try this or can we try that and. 297 00:16:50,760 --> 00:16:53,360 Do you, do you have an overall agent at ripling or? 298 00:16:53,520 --> 00:16:56,000 That's one of the new investments we are making. 299 00:16:56,800 --> 00:17:00,320 I think Mongo's partnership has helped us because if you think 300 00:17:00,320 --> 00:17:03,960 about Rippling, we are not just a payroll system, We are not 301 00:17:03,960 --> 00:17:06,599 just a finance or a travel expense system. 302 00:17:06,720 --> 00:17:12,560 We are a full suite of products. So for us, we can build an agent 303 00:17:12,560 --> 00:17:17,200 which is not a payroll agent or a finance agent or an IT agent. 304 00:17:17,200 --> 00:17:20,839 We can actually build a rippling intelligence agent which can 305 00:17:20,839 --> 00:17:24,400 actually answer questions across your entire day-to-day business. 306 00:17:24,680 --> 00:17:26,839 And that's essentially what we have embarked upon, right? 307 00:17:26,920 --> 00:17:29,560 Parker is extremely excited about business. 308 00:17:29,560 --> 00:17:32,800 Intelligence you think is a key output. 309 00:17:33,040 --> 00:17:35,400 Or I think it is really productivity. 310 00:17:35,880 --> 00:17:39,600 So if you think about AI productivity is the, is the 311 00:17:39,600 --> 00:17:42,560 impact which we we see in engineering or product and 312 00:17:42,560 --> 00:17:46,160 design or marketing, right. The same productivity impact is 313 00:17:46,160 --> 00:17:50,160 what we are hoping to provide to our customers because as our 314 00:17:50,160 --> 00:17:53,360 footprint has increased, our customers growth has also 315 00:17:53,360 --> 00:17:55,840 happened. So they are growing from like 316 00:17:55,840 --> 00:17:58,520 look at Andrew, one of our flagship customers. 317 00:17:58,960 --> 00:18:02,680 As Andrew is growing, they are using more products from we're 318 00:18:02,680 --> 00:18:05,080 playing, they have larger set of ibase. 319 00:18:05,600 --> 00:18:10,360 So essentially their day-to-day operations has gotten complex. 320 00:18:10,480 --> 00:18:12,760 Yeah. So now if you are able to offer 321 00:18:12,760 --> 00:18:18,120 them an agent to operate Android payroll to their IT to their 322 00:18:18,360 --> 00:18:22,720 travel and the agent takes care of, you know, proactively 323 00:18:22,960 --> 00:18:26,360 working with their administrators and making sure 324 00:18:26,360 --> 00:18:30,360 Android stays on top of things which they need to say on is the 325 00:18:30,360 --> 00:18:34,760 value we are aiming for. I hate to bring up a competitor, 326 00:18:34,760 --> 00:18:37,880 but ramp and rippling start in different places. 327 00:18:37,960 --> 00:18:41,400 They overlap on some. You now have, I think credit 328 00:18:41,400 --> 00:18:43,960 cards, yes. And you both want to be sort of 329 00:18:43,960 --> 00:18:47,760 the what, what's the term of art, like the record of all 330 00:18:47,800 --> 00:18:51,800 things business for somebody. They've clearly leaned into the 331 00:18:51,960 --> 00:18:54,960 AI brand more than you have. I don't know, what do you, what 332 00:18:54,960 --> 00:18:58,000 do you you're the AI guy? Like do you want, do you want 333 00:18:58,000 --> 00:18:59,560 that AI brand? Or it's like we have different 334 00:18:59,560 --> 00:19:01,200 sensibilities. Or how do you think about it? 335 00:19:01,400 --> 00:19:05,320 I think there is a, there is obviously an equity in terms of 336 00:19:05,760 --> 00:19:11,840 having a clear AI story around your company and your product 337 00:19:12,240 --> 00:19:15,560 because customers are keen to understand that because 338 00:19:15,560 --> 00:19:19,000 specifically because if I'm a ripping customer today or I'm a 339 00:19:19,000 --> 00:19:21,840 prospect evaluating Rippling today, I come with that 340 00:19:21,840 --> 00:19:25,840 expectation that there is a certain amount of AI automation 341 00:19:26,080 --> 00:19:32,080 productivity I will get. So it's, I, I commend Ramp for 342 00:19:32,080 --> 00:19:34,520 what they have been able to embark on, not just on the 343 00:19:34,520 --> 00:19:38,240 product side, but as well as in terms of positioning themselves 344 00:19:38,240 --> 00:19:41,520 as an AI forward company. I think Rippling is not very far 345 00:19:42,000 --> 00:19:44,800 in terms of our AI transformation internally. 346 00:19:45,560 --> 00:19:49,360 We are very further along than what people may be aware of in 347 00:19:49,360 --> 00:19:50,960 terms of our design partnerships. 348 00:19:50,960 --> 00:19:53,760 We have design partnerships with Cursor, with Open AI, with 349 00:19:53,760 --> 00:19:59,200 Entropic, with AWS where we get early access to the latest and 350 00:19:59,200 --> 00:20:03,360 greatest land chain data breaks early and even Mongo, so early 351 00:20:03,360 --> 00:20:05,440 access to their product capabilities. 352 00:20:05,680 --> 00:20:08,680 We are piloting that at tripling using that to create 353 00:20:08,680 --> 00:20:10,200 productivity and give feedback back. 354 00:20:10,600 --> 00:20:13,720 And when I see those interactions and when I see our. 355 00:20:14,160 --> 00:20:17,600 The ability to influence product road map of so many ecosystem of 356 00:20:17,600 --> 00:20:22,520 AI companies it reflects back on Rippling's AI impact is much 357 00:20:22,520 --> 00:20:24,840 larger than what people may be aware of. 358 00:20:24,880 --> 00:20:27,720 I should know that where is is enterprise search part of 359 00:20:27,720 --> 00:20:31,320 Rippling's vision because that's been an obviously clean is very 360 00:20:31,320 --> 00:20:36,760 promising AI company. I would say part of Rippling's 361 00:20:37,200 --> 00:20:40,520 AI assistant will take care of answering questions you'll need 362 00:20:40,560 --> 00:20:42,960 answers for right across your enterprise, right? 363 00:20:43,040 --> 00:20:47,600 Being able to obviously be being the system of record of a lot of 364 00:20:47,600 --> 00:20:50,240 information, right? You don't need glean in those 365 00:20:50,240 --> 00:20:52,520 cases because we can just answer those questions for you 366 00:20:52,880 --> 00:20:55,640 directly, right? What do you think? 367 00:20:55,880 --> 00:20:58,520 There's been a lot of AI is going to replace software 368 00:20:58,520 --> 00:21:01,760 companies. I mean, so you get it from both 369 00:21:01,760 --> 00:21:04,880 ends, embrace AI, but also like, oh, you know, you know, 370 00:21:04,880 --> 00:21:07,320 individual companies are just going to spin up their own 371 00:21:07,600 --> 00:21:11,000 payroll because, you know, they can build it with cursor on 372 00:21:11,000 --> 00:21:12,920 their own. Like, I don't know, payroll 373 00:21:12,920 --> 00:21:15,240 obviously to me seems sort of absurd because there's all the 374 00:21:15,240 --> 00:21:19,400 regulatory and it just like sensitive, but clearly there are 375 00:21:19,400 --> 00:21:21,960 there must be pieces of your business where people are sort 376 00:21:21,960 --> 00:21:24,000 of hacking together their own apps. 377 00:21:24,000 --> 00:21:26,880 Like, what do you make of this narrative that the rise of 378 00:21:26,880 --> 00:21:29,000 artificial intelligence is going to have all these sort of 379 00:21:29,280 --> 00:21:33,320 homespun software applications? We actually embrace that 380 00:21:33,320 --> 00:21:38,120 wholeheartedly because what we find at at Rippling is what we 381 00:21:38,120 --> 00:21:41,680 find at Rippling is as customers are using our platform, there 382 00:21:41,680 --> 00:21:47,760 are always niche unique use cases for which they like parts 383 00:21:47,760 --> 00:21:51,080 of what we offer to them, but they want to extend, enhance and 384 00:21:51,440 --> 00:21:57,320 add new capabilities on top of. Which is why beginning of last 385 00:21:57,320 --> 00:22:00,800 year, middle of last year, we launched our custom app 386 00:22:01,320 --> 00:22:04,280 capabilities where you can essentially vibe code an 387 00:22:04,280 --> 00:22:07,560 extension app to Rippling ad Rippling's platform itself. 388 00:22:07,600 --> 00:22:10,360 Interesting. And again, going back to since 389 00:22:10,360 --> 00:22:14,640 we are at Mongo's conference, right, being on Mongo in terms 390 00:22:14,640 --> 00:22:19,800 of ability to store generic artifacts and documents actually 391 00:22:19,800 --> 00:22:23,200 opened up that possibility for us essentially then to create 392 00:22:23,920 --> 00:22:28,040 custom objects which then can be used as a container for any 393 00:22:28,040 --> 00:22:30,040 customer to bring in whatever data they want to bring. 394 00:22:30,040 --> 00:22:34,080 In any interesting examples? That you get, oh, tons of very 395 00:22:34,080 --> 00:22:39,200 unique, very interesting examples in in companies, 396 00:22:39,200 --> 00:22:42,360 because we deal with a lot of workers who are dealing with 397 00:22:42,360 --> 00:22:45,840 shifts, for example, and compliance comes up quite often 398 00:22:45,840 --> 00:22:49,520 in terms of having them having gone through certain amount of 399 00:22:49,520 --> 00:22:53,120 rigors in certain industries. Very unique cases in terms of 400 00:22:53,120 --> 00:22:59,400 tracking, you know, ticketing, tracking, sliding sign outs, 401 00:22:59,400 --> 00:23:02,800 tracking. So there are like niche sort of 402 00:23:02,920 --> 00:23:05,800 use cases people are building, which has actually motivated 403 00:23:05,800 --> 00:23:11,240 Rippling to even launch RFD team, which is actually now 404 00:23:11,240 --> 00:23:14,960 going in to our customers and building these niche customer 405 00:23:14,960 --> 00:23:18,000 apps for them. And then seeing how we bring 406 00:23:18,000 --> 00:23:20,920 that into our platform as capabilities in terms of 407 00:23:20,920 --> 00:23:23,240 enhancing a platform to make that easier going forward. 408 00:23:24,680 --> 00:23:29,120 Do you think your code base gets a little worse the more people 409 00:23:29,120 --> 00:23:34,520 use AI tools or what's the risk in terms of over reliance on 410 00:23:35,160 --> 00:23:36,360 cursor? That's a great. 411 00:23:36,400 --> 00:23:40,640 That's a great question. I think we've started to observe 412 00:23:41,160 --> 00:23:48,080 that, yes, there is a dimension of an AI slop seeping in to the 413 00:23:48,080 --> 00:23:51,160 engineering discipline of coding every day, right? 414 00:23:51,560 --> 00:23:56,040 But I think we've been very strict about our AI stance on 415 00:23:56,160 --> 00:23:58,880 couple of things. First, testing is non 416 00:23:58,880 --> 00:24:00,800 negotiable. So any code which is being 417 00:24:00,840 --> 00:24:04,160 checked into the main line and pushed to production has to be 418 00:24:04,160 --> 00:24:08,480 fully tested and vetted. Second, accountability doesn't 419 00:24:08,480 --> 00:24:11,680 just the fact that I used cursor in AI doesn't take away my 420 00:24:11,680 --> 00:24:13,760 accountability of what that code does. 421 00:24:14,600 --> 00:24:17,920 And I think we've been very clear to engineers from day one 422 00:24:17,920 --> 00:24:19,680 that accountability is still there. 423 00:24:20,000 --> 00:24:24,680 So that essentially creates sort of a responsibility 424 00:24:25,480 --> 00:24:27,800 accountability in engineers to be cautious of it. 425 00:24:28,120 --> 00:24:30,000 But that's just on the human side. 426 00:24:30,560 --> 00:24:34,360 From a system side, we're also bringing in additional AI tools, 427 00:24:34,360 --> 00:24:38,760 so AI coding, code review tools, AI tools around looking through 428 00:24:38,760 --> 00:24:41,520 in terms of production outages and bringing that information 429 00:24:41,520 --> 00:24:46,000 back to to engineers to be able to troubleshoot and then 430 00:24:46,000 --> 00:24:48,160 improve. Because at the end of the day, 431 00:24:49,280 --> 00:24:54,080 the AI enablement or AI use is not just about development. 432 00:24:54,520 --> 00:24:57,320 AI can also be used in analysis of what's happening in 433 00:24:57,320 --> 00:24:59,360 production. AI can also be used in reviewing 434 00:24:59,360 --> 00:25:02,240 and verifying. And essentially by bringing 435 00:25:02,240 --> 00:25:07,800 those systems in those places, overall keeps the health of the 436 00:25:07,800 --> 00:25:13,600 code base at a peak and avoids the slops sort of just bleeding 437 00:25:13,600 --> 00:25:17,480 into production. In terms of the rippling 438 00:25:17,480 --> 00:25:22,520 product, maybe have you started to build it at all with the idea 439 00:25:22,520 --> 00:25:25,640 that it's like someday soon people are going to have AI 440 00:25:25,640 --> 00:25:28,720 agents that they treat as a worker? 441 00:25:28,720 --> 00:25:31,680 It's like, yeah, I want to in some ways compare apples to 442 00:25:31,680 --> 00:25:34,520 apple, this human and this AI agent. 443 00:25:34,520 --> 00:25:37,760 And I'm going to like think about it in terms of identity in 444 00:25:37,760 --> 00:25:40,320 the same way I would use rippling is that started to 445 00:25:40,320 --> 00:25:41,920 creep into. Your great. 446 00:25:42,320 --> 00:25:45,280 I don't know, it's like you're just just sort of picking up 447 00:25:45,280 --> 00:25:47,840 things we are discussing internally because one of the 448 00:25:47,840 --> 00:25:51,960 things we've been thinking about is agent identity, right? 449 00:25:51,960 --> 00:25:56,080 Because now if we think about agents we are building for our 450 00:25:56,080 --> 00:25:58,840 customers, at the end of the day, they'll start accessing 451 00:25:58,840 --> 00:26:01,040 payroll information. They'll start accessing, you 452 00:26:01,040 --> 00:26:04,320 know, information which has its own access management of who can 453 00:26:04,320 --> 00:26:06,560 see what and how and what type of actions they can take. 454 00:26:06,560 --> 00:26:11,120 Yeah, So it does need three things. 455 00:26:11,120 --> 00:26:14,920 One, obviously an identity of an agent, which we know who's doing 456 00:26:14,920 --> 00:26:18,720 what. But then second, it should still 457 00:26:18,720 --> 00:26:23,680 inherit from a human, because in some ways that's one of the 458 00:26:23,680 --> 00:26:26,560 conclusions. At least we landed on that. 459 00:26:26,880 --> 00:26:30,120 At the end of the day, humans. Culpable for this agent? 460 00:26:30,120 --> 00:26:35,240 Exactly, exactly, exactly who is who is on the hook for the agent 461 00:26:35,240 --> 00:26:39,080 at the end of the day and whose, whose team This essentially 462 00:26:39,080 --> 00:26:42,080 which translates into whose permissions, whose access 463 00:26:42,080 --> 00:26:43,960 management this agent has to have. 464 00:26:44,480 --> 00:26:48,280 And we found that it's a lot more easier to say that, OK, 465 00:26:48,280 --> 00:26:50,920 there is always a human. There is always a human in the 466 00:26:50,920 --> 00:26:53,000 system. Of course, what we are adding is 467 00:26:53,000 --> 00:26:56,080 a productivity element. So instead of in future thinking 468 00:26:56,080 --> 00:26:58,920 of adding more humans, you're essentially adding more agents. 469 00:26:59,240 --> 00:27:02,760 But they still function as part of a payroll team and then they 470 00:27:02,760 --> 00:27:05,680 get access according to being a payroll admin, right? 471 00:27:05,720 --> 00:27:09,040 They are part of an IT team and they get access as part of an. 472 00:27:09,040 --> 00:27:11,760 I feel like a lot of people are excited about what agents can 473 00:27:11,760 --> 00:27:15,880 do, but not excited about being the one responsible for stopping 474 00:27:15,880 --> 00:27:17,560 the agent from doing what it's not supposed to do. 475 00:27:17,560 --> 00:27:20,360 Is Rippling willing to take that on at all? 476 00:27:20,360 --> 00:27:23,600 Or like, do you, you see a world where you say, yes, we're the 477 00:27:23,600 --> 00:27:27,200 barrier for your agent not getting access to some system 478 00:27:27,200 --> 00:27:28,960 it's not supposed to? Have that already. 479 00:27:29,880 --> 00:27:32,200 That already is something part of Rippling's ethos. 480 00:27:32,600 --> 00:27:36,880 So we because we're not just a payroll company, we also have a 481 00:27:36,880 --> 00:27:40,120 full blown identity system, we have a full blown access 482 00:27:40,120 --> 00:27:43,360 management system. We already control what data who 483 00:27:43,360 --> 00:27:46,280 has access to. We have a very well defined 484 00:27:46,320 --> 00:27:48,840 access management platform essentially. 485 00:27:49,200 --> 00:27:52,800 And if you which is what is allowing us to now bring in 486 00:27:52,800 --> 00:27:55,640 agents and say, oh, this agent inherits from Encore and 487 00:27:55,640 --> 00:27:58,120 whatever Encore can do in rippling this agent can do 488 00:27:58,360 --> 00:28:01,640 right. And that philosophy essentially 489 00:28:01,640 --> 00:28:05,520 allows us to guardrail not just when we are building agents 490 00:28:05,520 --> 00:28:08,840 ourselves, but customers deploying agents in their 491 00:28:08,840 --> 00:28:13,200 landscape, being able to then start performing tasks and 492 00:28:13,200 --> 00:28:17,320 actions and leading to unintended consequences because 493 00:28:17,320 --> 00:28:20,480 we will still keep that principle even for them 494 00:28:20,480 --> 00:28:23,320 accessing ripping data. We guardrail that heavily. 495 00:28:23,480 --> 00:28:27,760 The API access is only enabled through certain controls and 496 00:28:27,760 --> 00:28:31,280 access management principles, so randomly customers can't deploy 497 00:28:31,280 --> 00:28:33,360 an agent and start accessing ripping information. 498 00:28:33,520 --> 00:28:37,080 Well, you have what every AI company wants, which is great 499 00:28:37,080 --> 00:28:38,720 data. And now it sounds like you've 500 00:28:38,720 --> 00:28:42,920 got the AI conviction. So exciting times ahead at 501 00:28:42,920 --> 00:28:44,600 Rippling and thanks for joining us. 502 00:28:44,600 --> 00:28:48,160 No, thanks for the great conversation and really enjoyed 503 00:28:48,160 --> 00:28:50,120 talking to you. And yeah, let's continue this 504 00:28:50,120 --> 00:28:51,040 dialogue. Sounds good. 505 00:28:51,040 --> 00:28:52,160 Thanks so much. Thank you. 506 00:28:52,440 --> 00:28:54,720 Thank you for tuning in to this week's episode of the podcast. 507 00:28:54,800 --> 00:28:56,640 If you're new here, please like and subscribe. 508 00:28:56,640 --> 00:28:59,240 It really helps out the channel. Listen in for new episodes every 509 00:28:59,240 --> 00:29:00,360 week wherever you get your podcast. 510 00:29:00,440 --> 00:29:00,680 Yes.