Ep # 177: AI for Post-Sales: The Quiet Effort Behind Data Modeling and Context ft. Jon Finegold (CEO, Precog)

Jon Finegold 00:05 Not only do they have access to the right information, but that they understand that information. And so, bringing in raw data on its own isn't enough. You have to bring in raw data with semantic context so that the models know exactly where to look to write SQL under the hood. Intro 00:20 Welcome to Across the Funnel, where we dig into concrete Go-To-Market moves across sales, customer success, and account management so you can build revenue that lasts. Brought to you by Hyperengage and Dextego. Adil Saleh 00:35 Hey, greetings everybody, this is Adil, Across the Funnel. This year has been so exciting in terms of the new age technology that we are talking about, and the founders and these GTM leaders that we are having these conversations about this AI bubble, the big, It's a big noise. And a lot of these tooling are trying one way or another to implement AI into their processes or their products internally and then for their customers. And we were thinking that why not we just, because data becomes a big ask and how to model, how to structure data in the way that it understands all these teaming and is in the best shape to be piped and populated across different tech stacks, be it engineering, be it product teams and all. So today we're gonna be talking to the Chief Executive of Precog, Jon. They're one of the fastest growing data infrastructure companies that integrates with all of your data tools to help structure data in the way that it is available in one of the bigger data platforms like Snowflake and Databricks. So we'll get to know more about this. Thank you very much, Jon, for taking the time. Jon Finegold 01:45 Yeah. Thanks. Nice to be here. Adil Saleh 01:48 So I know that your prior background is something relatable to this. I know there's no way you can be in a C-suite for a company like Precog without having a prior background. Could you just walk us through a little bit about your journey? For the last 10, 15 years and how you ended up, you recently, congratulations that you started on new role of Chief Executive at Precog, but prior to that, how this all came together and what was the biggest AI shift that you'd seen. And when I talk about AI, data is the bigger part of it, so how was your journey so far in the recent years? Jon Finegold 02:26 Yeah, so I'll give you just a quick overview on my journey. So I actually started my career as a software engineer, graduated with math, computer science, and I wrote code for about seven years and was involved with the early stage startup that grew very quickly and got acquired and kind of through that, I got very excited about more of the business side of technology, so took the opportunity to go to business school. Graduated from MIT Sloan in 1999 and was very early in B2B SaaS, helped launch OpenAir. I was the first employee hired by the founders, and it wasn't even called SaaS back then, but we sort of charted a course there in delivering software as a service. OpenAir was later acquired by NetSuite, and ultimately Oracle remains a leader in the category today. But through that, I got just really excited about just SaaS and cloud technology in general, and have been involved in a bunch of SaaS businesses over the last 10, 15 years. My most recent journey was at a company called Signiant, which was also in the data movement space, more focused on moving large video content for the media and entertainment industry. We grew that company from about 5 million in recurring revenue up to 50 million and sold it to private equity in the end of 2024. I was the chief marketing officer there, and so just kind of led the scaling and growth from the marketing side. And then after that acquisition I was out networking and connected with the guys from Venture Guide, who had invested in Precog. And Precog has some really, really powerful technology, has some really good customers, but not unlike a lot of startups, was probably trying to do too many things at once and was having trouble scaling. And so they brought me in to just kind of help take Precog to the next level. So I joined in October as CEO, and we're continuing to grow the business and really excited about the market opportunity in front of us. Adil Saleh 04:28 Absolutely. So now, when we talk about AI and when we talk about data infrastructure and data modeling, that becomes a big, big question. Like, how do you think, this AI has been now penetrating more towards less AI-enabled industries where these data warehouses are mostly used, like big enterprises in legal and healthcare. And then it becomes a big opportunity for people like you to have a huge impact with your expertise and all. So how do you think that this industry has been more moved by AI in the recent times compared to the beginning of ChatGPT, like 4, 4.5, when they were like tooling of GTMs, like customer support, sales, marketing, SDRs. There were so many AI agents coming through, and then now it's more like the heavy lifting of the AI, which is these industries. So how do you see these industry verticals changing with the impact of AI? Jon Finegold 05:31 We're seeing a big shift, I think right now, sort of end of 2025, early 2026. So we all know the power of AI. We all use it every day in our lives, in our work. But in the enterprise, when you think about asking sort of business critical questions. For example, which of my customers are at highest risk of churn? Or which of my suppliers are putting my compliance at risk? Or which of my SKUs are gonna run out of inventory in the next 30 days? Those are questions that the LLMs on their own can't answer because they don't have access to the data. So Precog is an enabler of that. And it's a really challenging problem because if you think about the LLMs that we use, when you go into ChatGPT, it's read everything that's ever been published on the internet, but it doesn't have access to your SAP data or your NetSuite data. And you can't just take that data and dump it into a model. For one thing, it would be prohibitively expensive. It changes all the time, so you need to be up to date. And it would unearth a lot of privacy, security concerns. So how do you solve that problem? How do you enable an enterprise to be able to answer those business critical questions that require access to very specific data sets that aren't available to the models? And so that's what Precog is all about. We have a platform, it's an AI native platform. We can automatically connect to any SaaS API: Salesforce, Marketo, you name it. We have trained it on hundreds and hundreds of APIs, and it can automatically spin up a connector. Not only extract the data, but understand the data and land it in your warehouse, whether that's Snowflake or Databricks or BigQuery or Fabric, whatever you're using, land it in a way that's not only normalized and in relational tables, but also with semantic context so that the models can understand that data and start answering questions right away. So we explain the data Adil Saleh 7:32 and the capability. Jon Finegold 7:34 We explain the data to the models, but we're only bringing in specific data sets that you want to bring in. We're not bringing that data into the model, we're just bringing the metadata into the model so it knows where to look, to basically under the hood generate SQL statements against your data in your data warehouse. But you can prompt it with natural language queries because we've explained, hey, this data, this piece of data sits in this table with this column name. And you do that across dozens of applications, tens of thousands of columns, it's a very challenging problem, but the more you can bring the data with the context, the more these models can really start to answer business critical questions. So it's a really exciting time, because I think we all see the promise of AI, but it hasn't been fulfilled in the enterprise because the models don't have access to the data. And there's a lot of obstacles in getting it there and getting it ready to use. Adil Saleh 08:34 Absolutely. And as you mentioned, the capabilities keep on improving, the capabilities and fine-tuning the internal data by the context that is fed in the backend. Jon Finegold 08:43 The power of the models is not the problem. We always like to say, AI is ready, your data isn't, right? That's sort of our tagline because we know the models are powerful enough to answer questions, but they can only do that if not only do they have access to the right information, but that they understand that information. And so, bringing in raw data on its own isn't enough. You have to bring in raw data with semantic context so that the models know exactly where to look to write SQL under the hood. Adil Saleh 09:12 Love that. And this is where all of these market leaders are doing really, really good, like feeding the right context and understanding the semantics of the context to have specialized capabilities. That's also the agentic workflow. Like agents are just like this. You might be noticing a lot, Claude bots and all these OpenCrawl and all these capabilities and skills inside Claude that are specialized for certain context. And now they have a longer reasoning or context window for you to use it as models and builds things on top. So how do you see this category, it’s more relatable to core B2B SaaS? Jon Finegold 09:55 Yeah, I mean, I think every business relies on data and uses lots of different applications. So we're a SaaS business, but we use a lot of SaaS applications. So even to run our own business, we're using, extracting data from those applications so that we can ask questions about customer churn and ARR and customer support situations and things like that. So every business today is a data business to some extent, right? We all rely on these apps and we all rely on the mission critical data that lives in those apps. And so making that available to the models is what unlocks incredible efficiencies, right? I mean, if you think about that analogy, like people started using ChatGPT and Claude several years ago. Well, now with Claude being integrated right into your GitHub repository, now it sees your proprietary code. The exponential benefit from that is remarkable. Well, think about now taking that same approach and applying it to your SAP data and your Oracle data, and your Salesforce data, and your NetSuite data, right? Or whatever systems you use internally, all of a sudden you can not just improve your software engineering practices with AI, but now you can improve your business processes, your regulatory reporting, your financial reporting, your customer 360 analysis. Like all of that stuff is possible if the models have access to the right data with the right context. Adil Saleh 11:26 Right context. That's super important. So now do you also see this category opening up towards not SMB, but mid-market, these smaller companies started to care more about their data models and how it's structured and how it's getting articulated across different functions and of course, like how it relates and impacts the revenue directly, which is retention and expansion models. So how do you see this category opening up towards the mid-market set? Jon Finegold 11:58 Yeah, I think the opportunity is there for any size business. I think it's just really a question of resources. So if a mid-size company has a data team, even if it's a small data team, but they need to bring data in from multiple applications, then the solution could be very beneficial to them. But that's really the gating factor, is like, do they have data in multiple places and do they have a data team that's focused on bringing that data together, governing it, and making it available to these various use cases. Adil Saleh 12:33 Mm-hmm. Jon Finegold 12:33 It depends. But like, my previous company Signiant, we had a small data team. We were bringing in data from multiple applications. We never connected it to AI at that point because it wasn't ready yet. But I could see, when I was CMO at Signiant, I would've loved to have LLM agentic access to all my sales and marketing systems. That would've been incredibly powerful. Adil Saleh 12:58 Yeah, I mean, it's just incredible that you look back and you think that, hey, this was just like a year back, could have moved things so fast. Jon Finegold 13:08 Yeah. I think about all the time we spent trying to understand our data and putting reports together and all of that. And again, today we could, using Precog and something like Snowflake and Claude, we could have enabled instant access, answers to those questions. Adil Saleh 13:26 Okay, perfect. So now let's talk a little bit about the Go-To-Market. I know this podcast was more ingrained towards it, but it was important for us and everybody listening to understand the unique positioning of Precog and what kind of impact that it's creating, especially in the mid-market segment, because most of these companies listening would be mid-market size companies, like between, I would say series A to series C. So, 10, 15 years old. So now thinking about Go-To-Market, how the customer education part is being the biggest concern, and how you're educating people or these personas to make sure that they understand the need of this and it keeps on growing as they grow into bigger companies. Jon Finegold 14:18 I think the good news is there's a lot of downward pressure from boards, investors, executives, to leverage AI in the enterprise, right? So everyone's thinking about it and I think just like I was mentioning, sort of end of '25, beginning of '26, you're starting to see people understand like, okay, the opportunity is really exciting, but there's some real obstacles in getting there. And so we're leveraging that momentum that's in the market to explain, hey, yeah, this is a real problem and this is where Precog can help. I really look at the evolution of data teams. Right now data teams, they kind of serve a bunch of folks at a bunch of different business units. They're essentially responding to JIRA tickets or just an onslaught of requests. I need this report, I need this report, I need this report. And they end up in kind of a spin cycle as a service provider to the business. I think AI has the potential to really shift that to where the data team becomes the true AI enabler for the enterprise. And I think the data teams that embrace that and say, hey, our job isn't just to build a report and respond to specific inquiries and tickets. Our job is to make sure that executives across the org can get the answers to the questions that they need, and that those teams become the true AI enablers. And I think that's super exciting. I like to say, the data teams can either be the heroes or the villains in this AI story, but the opportunity is there for them to be the hero. Adil Saleh 15:49 Are they ready to go outside their comfort? Jon Finegold 15:52 Yeah, I think it's a shift, right? Because again, the last five, 10 years, it's really been, hey, I need this report, I need this dashboard. And they're just in this perpetual spin cycle. So it is a little bit of a mindset shift, but they have the skillset and the knowledge to be able to do this. And we're starting to see some forward-thinking data teams lean in and say, hey, we can do a lot more if we leverage AI and we enable the right data sets to the right teams. But there's real challenges, right? There's privacy and security and governance and data cleanliness. I think AI is gonna expose some bad data and bad data practices. And so I think that's where the data teams could become the villain. But again, if done correctly, they could really be the hero in this story. Adil Saleh 16:44 So I was looking at here some of how you're packaging it across, especially mid-market. And looking at the success side of things, I know that this is more of a question for your VP of sales or account management or success, but on a high level, how do you have any kind of playbook to measure success across these companies and how they're growing, how they're leveraging your tooling based on connections or based on some of the metrics that they have to make sure that hey, they're potentially using the platform, adopting the platform. And at some point they'll grow enough for us to have some sort of expansion engagement and see how we can expand them. Or there's a churn indicator like they're not. Jon Finegold 17:29 Yeah, so we are a relatively small company, about 30 people. And we do have a customer success program and we have a sales team and we do use a lot of systems, from Intercom for tickets and Linear for internal project tracking, to obviously our own platform has data that we access to understand how much data is being moved, how many rows, how many columns, how many connections. Are they, is the customer doing more? Are they doing less? All of those signals help us know like when to engage and is this customer at risk of churn, is this customer an opportunity to grow. So we try to use those internally. We use the data and we use our own product internally to try and answer those questions. But that's how we model it, is based on, we know when we engage with the customer, we understand what business outcomes they're looking for, and we certainly try to track against those, but we also have just, like, and I would call them plumbing metrics of how many connections, how many different sources are they connected to, how much data is moving, is it changing, is it going up, is it going down. And those are good indicators of, hey, their data business is growing. Let's engage and see if we can help them even more. Adil Saleh 18:49 All of this is basically measured in digital touch motion, like in terms, it's sort of data-driven, like you get nudges or figures, your CS team and or with the next best course of action? Jon Finegold 18:59 Yeah, exactly. We're starting to use our own product and we've connected Claude by MCP to our Snowflake instance. So we're bringing all the data in from all of our systems into Snowflake. And then we've created views in Snowflake that have, like a customer success view and a sales view. And then we expose each of those views via MCP. And now I can connect Claude directly to that MCP server and I can ask questions like, are there any customers this quarter who are at risk of churn? And it will look at all that data and say, oh, here's the three that you should be proactively engaging with. Adil Saleh 19:37 Gotcha. Right. So the Claude MCP direct workflow integrated and based on analyzing all those... Jon Finegold 19:43 Yeah. And we happen to use Snowflake and Claude, but you could use any data warehouse platform and you could use any LLM, as long as it supports MCP, and you could run these workflows. Adil Saleh 19:55 Gotcha. So now I know that we are pretty much done with the first quarter, almost like a month in 2026. What do you think makes you excited in your category? Any kind of initiatives that you're thinking of taking as you've recently joined? Anything that is sitting on your plate for you to just execute this year? What makes you excited? Jon Finegold 20:17 Yeah, I mean, I just think the tailwind of AI and data in general is really good for our business. So I'm very excited about that. We are seeing now, I think this is the first year that I've seen real top-down pressure from executives and boards to start to take advantage of AI. And we are seeing that as a driver of our business, so that's great. I would say it's also a really crowded, noisy space. There's lots of tools and lots of people using similar words that we use. We have a really unique approach and a really strong competitive advantage. But thinking about how to break through that noise and educate the market on why Precog is a mission critical tool to enable AI-ready data and these agentic workflows, that's sort of the excitement and the challenge all in one. Adil Saleh 21:09 Yeah, I mean, with the challenge, with the adversity, there's an opportunity for growth and with people like you, with this massive background and all, been there, done that part of the story is always good to have on the team. So I wish you good luck with all of this and all the decision making becomes certainly tricky in certain markets, but it's always experimenting and failing faster and failing forward. And then you keep on iterating. So how big is the team alongside you, that is directly reporting to you? Is there any CMO or? Jon Finegold 21:45 Yeah, so there's about 30 people in the company, obviously pretty engineering heavy. So I think roughly 17 or 18 in engineering. And then we have a small sales team of about five people. I have a marketing person, customer success, partnerships. So there's a good size Go-To-Market team to bring our tech to market. But we are still in those relatively early stages. So we are more engineering heavy right now as we innovate and continue to improve the product. And as we grow, we'll start to scale up the Go-To-Market a little bit more. But yeah, I have a head of marketing, head of partnerships, head of customer success, and then a small sales team. Adil Saleh 22:24 I spoke to this one of the products in this category quite some time ago, like about a year ago, and it seems like a pretty long time. So I was just wondering, what are the multiples in your category in terms of when you're going about raising capital and thinking about valuation? Is it close to 5? Because of course, this is one of the unique categories that have better multiples than the other categories, especially marketing tooling. Jon Finegold 22:51 Yeah, it's a good question. I don't really know the answer to that because remember, I'm not the founder. I was brought in by the investors. So that round was already closed and I didn't go through the fundraising process. We'll probably go out and raise another round later this year. So I'll get a little bit of a sense. But I do think data and AI multiples tend to be greater than 5x. And we'll see what the market bears, obviously, as you know, it changes all the time. But yeah, we certainly see 10x multiples and even beyond in the AI space, and even in the data space. So, who knows? Adil Saleh 23:31 Yeah, perfect. One of our community members, they're also a SaaS company. We collect questions before we onboard. So are you open to advising these startups with some of the really cool founders with this huge background, they're building AI technologies and of course with AI, with data, but it becomes pretty lucrative in the beginning for a founder: hey, we are solving this, we are using LLMs and building wrappers on top to get the funding, initial round, but then when it comes to optimizing costs and having all the data infrastructure and everything they need, like people that have done it. So are you open to advising, consulting or anything for our community if anybody wants to? Jon Finegold 24:12 I mean, I'll be honest, I've got my hands full at the moment, having just come in as CEO and I'm laser focused on that. So not at the moment. I'm always happy to connect with folks and be helpful, but I don't think I have the bandwidth to be a true advisor at this moment in time. Adil Saleh 24:28 Yeah. Jon Finegold 24:29 Certainly glad to connect with folks and via LinkedIn and answer questions and that sort of thing. Adil Saleh 24:36 Love it. So one thing, another question coming from our team was more like, what culture shifts or initiatives that you're taking from your prior background, previous roles to this new role. Like anything that you think should be the operating principle or how to get the best work out of your people, team, leadership, all of that. So any new initiatives that you're taking? because people are people. Someone with your background... Jon Finegold 25:02 Yeah, I mean, I think one thing that we've put in place is to really be customer first. I know that sounds obvious, but just making sure that we are proactively engaged with our customers, talking to them, listening, learning. To me everything starts there. And not being technology first, but start with what are the customer problems? What are the pains, what are they feeling? And then working backwards from there. So that's something that we focus on a lot as a culture. And then also, we're at that point where the founders, brilliant, brought in a lot of creativity and incredible technology. One of the things that I've brought in is just a little more structure and how we measure and manage things, so that we can just understand where we're at as a business and understand when to grow certain dimensions of the business and that kind of thing. So a little, what I like to say is, enough process that's appropriate for our size and scale. Adil Saleh 26:04 Gotcha. So by that you mean the commercial side of things? The commercial view of a business? Jon Finegold 26:09 Yeah, just making sure that we're measuring all the work that we do and holding ourselves accountable to tasks and outcomes and that sort of thing. We're just at that stage now where we need a little more discipline. So that's part of growing. Adil Saleh 26:29 Yeah, that's more for the team and founders at Precog. What is for the audience, what is one thing that you want? I know a lot of these founders, they wake up and open LinkedIn every single morning and they see a huge stream of products getting funded. I mean, success stories, and nobody of course, nobody shares the failures. Everybody's sharing the success stories and wins and everything. So this creates a lot of, I would say not depression, but they feel like they've been left behind or they're just struggling with perseverance and they always want to do more. I know it's always good to stay focused and have your product vision aligned and everything. So what is one thing that you would advise as a leader yourself to keep up with this emotional baggage that they carry, looking at the external inflicted noise they get to experience? Jon Finegold 27:24 Yeah, I mean, again, that's a good question. I think the key to just keep the team energized and focused is, people get excited about seeing customers have success. Right? And so that customer first mentality, not just thinking it, but really living it and then making sure that everyone in the org, even if they're not customer facing, kind of sees exactly what's happening. So just having a more open flow of information from the customer-facing folks back down to the product team and making sure there's a good loop there. That to me is what energizes folks. That's why people join startups, they want to build something and be successful. And so seeing that, feeling it, embracing it, being a part of it is exciting. Adil Saleh 28:11 Yeah, being a part of something that's growing like a startup and making and feeling the impact that you're creating, either you're customer facing or customer support or product member of the team. Perfect. Jon, it was really, really nice meeting. It was insightful, getting to know you, your story. You carried a little bit of how Precog is making this huge impact and we only wish you good luck for the next round and all these initiatives that you're thinking of bringing on with this team. Jon Finegold 28:42 Thank you so much. It was a pleasure to speak with you and look forward to staying connected. Adil Saleh 28:46 Sure, why not? Outro 28:47 Thank you very much for listening to Across the Funnel. If you got one useful GTM idea out of this show today, please share this with a teammate and hit follow. Explore Hyperengage at hyperengage.io and Dextego at dextego.com.