CRM As A Business World Model: The Future For GTM Teams? | with Keith Peiris, CEO @ Lightfield

0:00Forecasting now, prediction now is really hard. People are putting out new products every couple of weeks. The model costs are dropping. Segments are being written and rewritten. That's Keith Peiris, co-founder and CEO of Lightfield, who just raised a whopping$47 million Series A led by Andreessen Horowitz. Why? He's building an agentic CRM already being used by more than 5 ,000 companies. In today's episode, he shares what makes an AI-native CRM powerful, but also how they're tackling the issue of accuracy and reliability for sellers. The models are always going to be non-deterministic. I see that deterministic problem as a harness problem, and let me explain.

0:39We also debate whether the SaaS partner ecosystem is dead. Keith shares the two factors that drive a company to rip out its CRM, and why his deals with 200-person companies land at values traditional SaaS only sees at the enterprise. Look at our deal size for a company that's 200 people. This actually looks like a deal size for Salesforce with like 5 ,000 people. If you're a revenue leader in the age of AI, this episode is for you. Welcome to Topline. Hey, everybody. Welcome to Topline. It's Sam Jacobs. I'm your co-host. I'm the CEO and founder of Pavilion. I'm joined by Asad Zaman, the CEO of Sales Talent Agency, and A.J.

1:16Bruno, the CEO of Quotapath, which is the leading designer of incentive compensation strategies and systems for AI native organizations and other organizations as well. All organizations, all of them, Sam. And we've got a very special guest on the show today. We're really excited to welcome Keith Paris. He's the co-founder and CEO of Lightfield. And Lightfield is an agentic CRM that ingests every customer interaction and structures that history for agents to work from. Lightfield launched in November 2025, so a very young company, but has already been adopted by more than 5 ,000 customers and has raised$47 million Series A, led by Andreessen Horowitz, just announced a couple of weeks ago.

1:59Before Lightfield, Keith spent more than a decade at Meta, where he grew Instagram Direct to 500 million monthly users. Keith, welcome to Topline. Excited to be here. we're really excited to have you so i gave a little bit of the preamble on just now but in your words tell us what is lightfield what do you do and i think the important thing for us all to jump into this conversation is why now why is now the time for a platform like lightfield and what are the systems that you look to displace or disrupt as you gain entry with uh with your with your products sam maybe i'll start with the why now and back into into lightfield which is that we were noticing how fast models were improving and advancing and taking in more information, reasoning through it.

2:48And we had the simple thought of, wait a minute, if these models are going to be two times as smart, three times as smart, as a great operator, what we need to build for them is a model of your business. And if we can build out a model of your business, then they'll be able to do great work, help you with hard forecasting, didn't help you with deep analysis. And we thought the best way to build a model of your business was to start with your CRM. You know, I think like every company's lifeblood are their customers and their prospects, what they say, what they do, what they actually committed to you.

3:25So we decided to sort of reimagine CRM, you know, in the context of building a business world model for a company. and CRM is a crowded space. You know, so one of the challenges that, I mean, it seems like you're already exploiting the opportunity. You've got 5 ,000 customers in just a few short months, which is incredible. So congratulations on that progress. But it is true that CRM and go-to-market generally is a very crowded space. What were the key insights? What are the things that you believe, Peter T. might call them your secrets, that gave you the confidence that now is the moment for this particular solution, this particular product.

4:07So I'll go back a little bit in time, which is to say that all of these systems of record, whether it was ERP or first CRM, they're actually all about forecasting. And I think the reason that Salesforce was able to break through was that sales reps didn't use Siebel. And because they didn't use Siebel, as a revenue leader, you couldn't trust the pipeline forecast, and someone needed to invent a tool that would gather data to make the forecasts more reliable. The thing that got me sort of excited about this phase is that forecasting now, prediction now, is really hard. You know, people are putting out new products every couple of weeks, the model costs are dropping, segments are being written and rewritten.

4:57And I think sort of forecast and prediction in today's world is far more complicated than just looking at your pipeline for the past month, the past quarter. So I had thought, you know, to really do forecasting in this world, you have to build something larger than a CRM. You know, you have to build, you have to get signals, you have to understand the world, you have to understand what's the company doing, what are they building. And that already felt pretty different than a sort of a, you know, a down the fairway CRM. So that felt different. And then the other thing that felt different to me was sort of the data structure required in order to build the CRM of the future, which felt so different than a HubSpot or Salesforce.

5:38You can now build a system that basically updates itself, sort of automatic by default through APIs and the conventional CRMs weren't built for that. And then two, I think you want to capture everything versus what's just in the fields, because who knows what question you're going to ask next month. And we just thought that those things combined were so different than sort of established CRM that there was a shot for someone to invent something new. Does that mean that the target market for this new generation of CRMs, URIs and DAE and all these other AI native CRMs, is that primarily an inside sales focused go-to-market team that would be the ICP for these?

6:27because obviously a lot of the value is in capturing everything, but you can't capture everything. You know, when it comes to enterprise sales, it's field sales. You had dinners and conferences, et cetera. It's kind of creepy to be like, let me record this dinner. So does that define your segment as like the SMB portion of go-to-market? It's a good question. I would say we have a lot of companies that sell to enterprise who use partial capture. You know, they record calls, they record in the behavior, or sorry, they record product behavior. They maybe use granola, you know, in their on-site meetings.

7:05And it gets them a little closer, to your point. You know, we don't have any customers wearing the pendant at dinners recording everything. But it gets them, you know, meaningfully closer than if they were using a traditional CRM. I don't sit next to people that have Apple Watches anymore at business tennis because I just feel like they're recording me in a creepy way. Yeah, we're not going to sit next to each other, Sam. Far, far. How do you feel about the cameras I had surreptitiously installed in your house? I know they made you happy, so I'm okay with that. Keith, first off, to Sam's point, congrats on the progress.

7:44but I mean, I guess 5 ,000, when did you launch officially? It was November last year. Okay, tell me more about 5 ,000 customers. Because listen, I mean, I've been around for eight years. You have like 5x the number of customers that I have. Tell me, what's your secret? How? Very generous accounting. So I think the honest answer is we started small, which is that we built the system for two years. And when we launched it, it was really only ready for, you know, a seed series A company. So we started with a PLG motion, get some feedback, really serve those customers well, and then move into, you know, a real sales led motion.

8:26So I would say a lot of those companies that we serve are startups. And then, you know, thankfully, a few of them grew. You know, actually, some of the startups we serve went from no AEs to 100 AEs. It's just the nature of the times. Right. and then we fought really hard to make them happy. And then we sort of used those reference points to sort of move into mid-market. I shouldn't say move in past tense, move in the present tense. We're sort of moving to the larger side of SMB mid-market. And we just leaned into being the best CRM when you don't have one. And sort of moving people off of the traditional SMB ones as their companies scale.

9:08So 5 ,000, it sounds like you learned a lot from those early smaller PLG customers. Do you want them to stick around or no? Are you actively like, look, we've learned a lot from them, but we really need to displace incumbent, whether it's HubSpot or Salesforce, which I'm curious if that's actually taking place, or if you're really focused on maybe monday.com CRM users that are needing to move off of it. What is that those early cohorts of customers feel and sound like? and do you actually want them as your customers today? Yeah, yeah. It's a good question. I would say that it's a really important lead generation tool to have this PLG product where you raise some seed capital, you need a simple system that collects everything, and then as soon as you hit product market fit, you need something real.

9:59So we actually moved quite a few people off of our 2K per year plan to the 24k one as soon as they hit PMF. So I think it's an important part of our acquisition motion. But right now, I would say that our ideal customer probably has 10 reps going on 30 as opposed to sort of wandering in the desert looking for product market sets. This is good. So now that we understand what you're doing, let's talk about the world that you're building in and the world we all operate in. And so the first thing I've been thinking about a lot, and this is in relation to what you're doing and what others in go-to-market tech are trying to do, there's this fundamental thing that I keep coming back to, which is the mistakes that the AI still meets.

10:52you know it is we're trying to do deterministic things with probabilistic technology so forecasting is like you know if you miss the forecast by five percent heads get chopped and five percent error rate is not you know unheard of for the top models that we have today and so it's like it almost feels like this problem that i don't know if it can even be solved because the technology is fundamentally going to have errors in it. It's going to record, capture information, but it'll make mistakes there. It will do analysis, but it will make mistakes then. All of those little, little mistakes compound into a big mistake.

11:35But then there's no way to get around that. There literally isn't. Technically, you can't get around it. So how do you deal with this? Is this just belief that one day we'll have a model that doesn't make mistakes? It's a good question. I think what we're building with the assumption that the models are always going to be non-deterministic. I think it's the nature of sort of an intelligent being. My team, as much as I would love for them to be deterministic, is sometimes a little unpredictable. They're brilliant folks. I see that deterministic problem as a harness problem. And let me explain. The first thing that we set up for most growth stage companies using Lightfield is we actually don't tell them to leave everything on automatic.

12:21We're like, if these fields are important, you should ask your reps to approve Lightfield's suggestions so that you have auditability. That's the first thing. If you don't have auditability of your fields, you'll never believe your forecast. and then you know for for the processes like pricing discounting uh qualification account research we find that you have to give the models quite a bit of context for them to to be reliable that way so but can i take into the forecasting part because like i i think qualification etc like we're willing to take a hit there right like if if you have 10 leads come in and the ai like misqualifies one like fuck it like had one useless conversation so what like you know you kiss a couple of costs.

13:06But forecast is very different. So then there becomes this human in the loop auditing the AI. But if you actually think about this big itself file that comes out with all these formulas, taking all this information in to give us an end number, now you have a human going and checking every formula, looking at every calculation, basically rebuilding manually the AI generated version of that thing, how does that save anyone a minute of time? Like, it almost sounds like more work to me this way. And this is not just what you're doing. This is like across the AI stack when it's happening in finance, it's happening in a lot of places where you do not have the tolerance for a mistake.

13:50You have humans that are just literally spending as much time as they used to checking the work of the AI and catching its mistakes. And there are tons of mistakes there. It's a good call out. So what I think about our dashboards, which are obviously in our tool, we did a lot of implementation work up front. You know, my head of finance looked at every formula. We converted some things from prompts to real code, right? I don't want LLMs doing math. I want real code doing math. And that got us, you know, 95 % of the way there. And Now we just make fine adjustments week over week. So I think it's really about figuring out what's the right tool for the job, which is to say that I don't think LLMs are the right tool for math.

14:36Some might disagree with me, but that's my current take. But I think LLMs are great at sort of finding insights. So for me, the influence to the forecast is like, wait a minute, yes, we closed this many deals last quarter, but how many of them were sort of on the cusp of actually being able to close because of missing features, your competition and so forth. And I think that's sort of the value of the LLMs to the forecast. Does it play a role in like deal reviews? Like it's one of the companies spend so much time in deal reviews and there's such a variance, you know, there's great deal reviews and then there's like the average deal review meeting and so much gets lost.

15:19Like there's so much nuance that you're trying to catch in a deal review meeting that usually gets missed. Like, do you see this sort of technology capturing, like, let's say there's a meeting and it's recording the meeting and it can tell that the rep is showing signs of nervousness when giving, you know, their take on a particular deal and it's looking at signal. Like, do you think your technology will play a role in helping deal reviews get better so that forecasting at the end can get better? I think so, but I'll add a caveat, which is that I think a great revenue leader has a very strong opinion on how to run a deal review.

15:58They have a strong opinion on what to pay attention to, what they care about, what they want to push the reps on. and I think you can't leave that up to chance. I don't think you can even give that to GPT-8 and be like, you know, look at my company, see what happens. I think you have to sort of set the standards for how you want deal review to run and then you can get the model to follow, you know, your skills, knowledge, automations on the way you want to run your team. AJ, how did you run your deal reviews? I was literally just thinking about this. What would be the compelling event for my team that uses Salesforce to switch to Lightfield?

16:34and would deal review would it be deal reviews would that be like enough as a part of it first off to answer your question i made my reps cry when i did deal reviews and pipe review that happened a lot you made a lot of people cry i know a lot the big softie right now is like not canceled on last door by now does anyone look at glass door anymore please don't look at my glass door i i'm what i'm thinking about and this is something that we're dealing with keith and And I'm curious where you have a quick, basically what you're saying is you have real data that has a query. If the query that data, then you have a semantic layer that sits over top of it.

17:11And that's what actually ultimately controls everything here. That gives you like your AI intelligence layer. And we have this, we do so many evals and it sounds like you spend a good amount of time your morning, like an hour a morning. you're in your own product and you're actually doing these type of evals so that you catch these things before your customer does right how often in the morning are you coming across these where you're like oh we need to like oh shoot i like this is a fire drill i have to go figure out what's broken right now because otherwise it's broken for all of our customers it's a good question it's very rare that we find something that is a a real error as in i wanted lightfield to do X and we get Y.

17:54I would say that that happens, you know, less than once a month. But the most common error I see running Lightfield, the company, is that I told Lightfield that I wanted X and then we learned from the field that we shouldn't be paying attention to X. We should be paying attention to like X plus one or Y. So what I thought I knew about our go-to-market process was, you know, proven wrong by something we learned recently. Let me go back to the other question that I really wanted to riff on a little bit. I'm going to ask this for Sam and Asit. For you, we're CEOs. You're pitching us. What would be the compelling event or why would we switch from Salesforce, HubSpot, whatever, to Lightfield?

18:36Or is your market somebody that doesn't have a CERN at all? Or are we not the ICP right now? You would be, AJ. You're a tech company. I think I am. Yeah, I think I am. Yeah. I can tell you, like, I'll give you my take on us because we're actually switching right now. We switched CRM slash ATS a lot in the last couple of years. And we're churning again. They don't know yet. STA as a customer, Quotapath knows this. You have had like six tech stack migrations. Yeah, Quotapath has to build integrations to every new thing. They're like, what is this nonsense? So much fun. But he loves me, so he does it.

19:16So we're switching again, AJ, be ready. it was basically it's actually not as practical of a reason it's when your reps start just vocalizing how the thing that they spend a lot of time in has zero roi for them because like the whole purpose of these technologies is to help leadership and management understand their businesses better so that they can more effectively allocate resources within that business for which the information has to be really good in it. And for that, the people have to get some sort of value for giving you that exchange, right? Like the person has to be like, I get all this value from it, so I'm gonna do these things for in it.

20:00And when you start hearing that there's just no value, you know that what you're gonna get in there is crap. And if it was purely inside sales, you could have a CRM that just records everything. But like when there's more, when you can't record everything, you actually do need your people to be engaged. And when I hear that, it makes me scared. The other one is you start like looking at it. This is, I was looking at it and I was like, this looks old now. It looks old, it feels old. It feels like I'm driving like a car from like - Salesforce has looked old for 25 years, my friend. I don't like it.

20:36I want something to look cool. Like this is like the thing you spend a lot of your time in. do you want to feel like you're driving an old car or you're driving a new car like there's that sort of thing like how does it make your people feel and you feel in terms of your business if you're spending so much time within it um and the last thing was like how well does it play with ai and what we had was not playing well well let me push you on the first one though because do you want your reps spending all their time in the crm today like i don't think so i want to see I want an ATS slash CRM that gives them value of some sort and mates them for which they spend time.

21:15I don't want them spending time filling forms out and putting information in it. But it has to be more than information gathering for the information that's being gathered to be high quality. Like it has to, there has to be a value exchange. So an example is in my world of search, a person should be able to take a job profile that they're working on, dump it into like the ATS system be like look at the entire database tell me who are the people that I should be talking to that we spoke to in the last three months that match this thing precisely and it should like this come up with something now it's a product for my people whereas like this thing can't do that and that's the sort of thing so once your people feel like they're getting value then they're willing to like eat shit and put the information in it.

21:58Do you agree with Asad's points there is there anything like that resonates well or maybe different than how you're thinking about with Lightfield? It absolutely resonates. It's funny, though. Now that I've done a few sort of rip and replace of Salesforce and HubSpot, it's funny. You'd think from the outside that better agentic work is a reason to change. And it's not. which is to say that I could put 10 emails in front of you. One could be marginally better, one could be marginally worse. It's not a good enough reason to change a system of record. I found that there are two motivating reasons for our customers.

22:38One is you have to believe that modeling your entire business makes every part of your revenue funnel better. Let me explain by that. if you believe your CRM knowing about your happiest customers and your most intense customers is going to help you find new customers, then that's a good reason to have it all in one place. So a lot of SDRs that use Lightfield, they're pulling real-time insights from customer success. They're pulling real-time case studies from their happiest customers. They're seeing renewal ahead of time for the expansion folks. You have to believe that that interconnectedness is worth something.

23:24And that's a good reason. And then the other reason that I've heard is actually this higher-level scenario planning, which is to say, next quarter, should we build this product feature? Should we invest in this segment? should we change our sales process you know and we found that for that those sorts of questions the modeling really matters the data completion really one thing that a lot of people are doing that i think is pretty dumb is telling the reps what to do next i think that is ridiculous that again it's trying to make it's a bunch of technical people that have convinced themselves that sales is deterministic right like say this do this say this do this then this will happen And it's like, okay, dude, when you sell, it's not like that.

24:13Like, especially when you sell anything with some level of complexity, the same move in scenario A versus scenario B leads to different results. It is probabilistic. And so when the AI is trying to tell salespeople that now it's your turn to say this, and now it's your turn to say that, I think those companies go down in flames. It is just a misunderstanding of the craft of sales. And maybe it's okay in customer support. but they should build a customer support platform. Then why are they building a sales platform? But they're like, we're building a sales platform and they're trying to build it in a way that you would build customer support.

24:49It's ridiculous. Keith, along those lines, how do you think about, you know, because I think one of the things you've mentioned in the past is focus, right? Depth of features versus breadth of features. And it is true that everybody's trying to be everything right now. and it makes it, you know, I used to joke or I don't know if it was a joke. I used to comment, Gong and Outreach used to be close partners back in the SaaS days. And, but everybody's, everybody's roadmap ends up in the same place. And when I was saying that I was only talking about a few features, but it made it so that Gong - HubSpot and Salesforce were partners.

25:29HubSpot had an office in Salesforce's building. And so exactly to the point, right? And so ultimately, SalesLoft, Clary, all of these pre-AI companies, they ended up building all the same stuff. And Gong rolls out forecasting, and I'm sure, and Clary has to merge with SalesLoft, and they roll out sequencing, etc, etc. And then you look at these AI native go-to-market companies and it's to the extreme. Because now it really isn't a meaningful lift the way it used to be to push out a platform that has a tremendous number of features. And so everybody has, you know, when you've got competitors, we're out of the box, they kind of do everything.

26:08They have a sequencing tool. They have a call recorder, of course, because everybody's got a call recorder. And they've got, you know, enrichment, waterfall enrichment for data. And you can build a lead list. So how do you think about, which is all a long-winded way of saying, first of all, you have a point of view on going deep versus wide. I'd love to hear more about that. And then on a broader level, how do you think about competition in this new world where features are just a really difficult place to, as you would say, build a moat, as we might say. So how do you think about breadth versus depth and in the larger competitive set?

26:47First, I would say, you know, for our ICP, and right now we're focused on the high growth tech companies, they all want best of breed. And I think if you're in a high growth company, you will want a best of breed solution for every part of your, you know, your revenue operations. There's a question of like, what does best of breed mean? You know, does best of breed mean every single feature in outreach for a sequencer? Or does best of breed mean, hey, because this system deeply understands my customers and my sales process, it's going to be able to do the best agentic work. Right. And it's actually sort of an analogy to, you know, Microsoft Office, right?

27:31It's hard to argue that Microsoft Teams is like the best messaging tool or, you know, that Microsoft PowerPoint is the best slides tool. But what makes them the best is actually that they're all connected to each other and they all share the same data, the same directory access. And that's what makes it work. And I think the bet that a lot of people are making now is actually if you have the best modeling of your customers and your prospects, that makes those other features better than the points. Can I ask you a question on this? Yeah. I've been thinking about this a lot. If you look at Salesforce as an example, a big part of their success was the app ecosystem that they built.

28:18So it was like this really good foundation on which you could go and build something on top of it. And all the value that their partners helped create through investments in R &D and innovation and building and go to market, the downstream benefit was on Salesforce. It created a dependency for those companies. That's why somebody that was built on top of Salesforce or Shopify always traded at a lower multiple than the company they were built on top of. But it was a really good way for the core business to expand the value that it can provide to the market. But at the time, building things was hard.

28:56And so it was like, I'm going to do this thing and you do these things and we're going to like connect the dots and it's going to be fantastic. Now building things are becoming easier. So all the companies, because what you're saying, a bunch of companies in other spaces are also saying that first you have to get the foundation right. And then you can attach other things to it. But you need standalone specialist things that can attach in, whereas all these other companies want to be everything. If everyone wants to be everything, how does the company that's building the foundation that needs other people to tap into it to create the maximum value, how does it actually do that without having to also then end up doing everything eventually?

29:36Yeah, it's a good question in the sense that I can't imagine us doing everything related to revenue. Like everything related to revenue is infinite, you know, and there are great companies that, you know, will endure. And I don't think that we will be able to get to their problems in 10 years. So I think it's important to draw that line. and I think the reason this line is blurry you know is that a lot of the like tech around go to market actually has become commoditized you know which is to say that you can build a good sequencer you know over a couple of weeks you know you can build good forecasting tools with gpt6 sitting on top of great data over a couple of weeks so I think that like that layer you know the sort of Clary, Outreach, like that layer, I think, will probably be sucked into the system of record.

30:34But there's so much beyond that, you know, that there's... That's a lot for you to build. You have to build play, Outreach, Gong, like you have to build all these things on top of what you're building. And you kind of have to build it really, really quickly, because by the time your Series A company gets to Series C, they need a real CRM, like a bid one. And so you need to get there by in the next two years, you have to have all those things that you build into this. And there might be like little bits and pieces that stand outside. But like, Monaco wants to do everything that you want to do, right?

31:09Like, so it's like, you actually do end up having to, but that was my point is that you can start by saying, let's get the base right. But eventually, you have to build most things yourself. Yeah, yeah. Oh, for sure. I would say it's definitely a race between us and the other startups. And it's just a question of, you know, who started in the right place? And because of where they started, who's best set up to build, you know, the next 10, 15, 30 things? If not, you'd be partner with AJ. Don't do it. Don't do it. It's too hard. It's too hard. There is this, like, funny thing that's worth saying.

31:46By no means do we have the best commission planning tool yet. But, you know, he did do this. By no means. By no means. Let's demo it. Let's demo it. Let's find out right now. I'll share my screen. No, my commission planning is not ready for primetime. But the funny thing is that a lot of these, I would say a lot of these features are now just automations and skills that run on your system of record in the new world. And I think that this is the crazy thing about building software. You know, we have a lot of our customers are building CPQ into Lightfield. You know, they've set up a bunch of automations.

32:33They've set up a bunch of conditional fields. They have forms that come up. CPQ is complicated. Didn't like Kleiner just invest in Roadrunner, which is like they intubated Roadrunner and Kleiner invested in New Rise. It was a really interesting investment. And like CPQ is like a really boring old space. And like, there is the need for something new over there. I think that's what Roadrunner was trying to do. It's hard to imagine somebody can vibe code a really good one themselves. There's so many like permissions and rules and this and that. I don't know. I have a friend that my friend Ash Lomberg at Chargewee just vibe coded as CPQ.

33:09This episode of Topline is brought to you by Compound Growth Marketing. The way your customers buy is shifted. Before they ever reach a rep, they're asking an LLM who to trust, who to shortlist, and why. Decision-making is impacted by investment at the top of the funnel, but most B2B marketing teams underinvest because it's difficult to measure. That's the gap, and it's expensive. Compound growth marketing bridges this gap. For many of the fastest-growing companies in B2B SaaS and cybersecurity, their Predictable Demand System, or PDS, unites Paid Media Answer Engine Optimization, or AEO, and GTM Engineering or GTME into one discipline engineered to produce pipeline you can actually forecast backed by over 100 million in media manage.

33:52No vanity metrics, no guesswork. If you'd like a free audit on your funnel, including AEO, reach out at compoundgrowthmarketing.com. That's compoundgrowthmarketing.com. And mention Pavilion. AJ, Asad, tell us about your reaction to what I just said. Love it. Compound Growth Marketing feels like it helps B2B companies. Turn growth from a collection of tactics into a repeatable system, Sam. Ho-ho. Ho-ho. A.J. is out there for compound growth marketing, slinging products, slinging SKUs. He wants to keep doing top line and he knows we need to pay for it.

34:31Asit, are you a fan of compound growth marketing? We're a fan. Anyone that helps cyber, such a commoditized space, if you can help those companies generate demand, you're probably very, very good. So worth looking at. Hey, this episode of Top Lines brought to you by our friends at DevRev. Listen, I want to tell you something I've been really excited about lately, and it is the brands that DevRev is building. One of them is called Computer. But it's not the computer you're using right now. It's a different computer. If you work in GTM or sales, you know what it's like. You're running a dozen deals at once.

35:01It's on you to remember every detail, what the customer said, what you promised, yada, yada, yada. So things slip. A follow-up gets missed. And you might think, I'm going to use my computer to fix that problem. but you're not going to use your computer. You're going to use Computer by DevRev. It's an AI teammate that fixes everything. Computer uses something called shared memory to hold the full context of every deal across all of your systems, even across the rest of your team. So nothing falls through the cracks and you always know exactly where every account stands. Take a look at computer.io forward slash top line, computer.io forward slash top line.

35:35Asin, you like sharing everything from all your disparate systems into one thing. called computer i think it's very useful that said i use everything on incognito mode so i'm probably not the memory guy sam i've said this before no i've seriously said this before but like an ai teammate that actually remembers what i'm doing and how my company works every day at every moment so i don't have to start from zero every single day like os it has to apparently is amazing so don't be os it don't ever be os it and codettes don't have incognito mode when you download the app it's only on the web version i find that very annoying i think you know on the one hand i think os it's clearly a very paranoid person but on the other hand based on my my knowledge of his business dealings i think it's appropriate anyway computer.io forward slash topline to buy their products hey everybody this episode of topline is brought to you by pavilion and pavilion's gtm event which as we record this is happening two weeks from today.

36:40Two weeks from today, we'll be in New York City with Asit and AJ. We'll be doing live top-line recordings featuring... What are you most excited about, Sam? Top three. Okay, that's great. Here's what, first of all, I believe, unlike my wedding, that this will be an event that you actually show up to. I will be. I promise. Let's start here. I'm excited to see you in person. You did RSVP yesterday, so we don't... I have to support my assumption here, except that we've booked a lot of great... Sam has sent out four live Topline events. I've RARSVP'd yes to all of them. Asad, have you? For what? For when we're doing the live Topline recording on Monday.

37:23I will be there. Yeah, I'm there. I'm in New York. Well, what else am I going to do? We're going to have Manny Medina, CEO paid. We're going to have Kiva Kolstein, president, CRO of AlphaSense, about to go public. We're going to have the guy from Anthropic, your friend, Asad. Peter Nolan Peter Nolan and then I think we might have Cassie Young as well who by the way is very very pregnant and expecting her first so that's really exciting Michael Larrant so I'm excited for the live in person that but I'm also excited for the Pavilion Gold dinner that we're having on Tuesday night you both will be there we will be there it's going to blow your mind AJ's been to a Pavilion Gold event before Asad you have not we do these things are over the top Jale will be there so we'll get to spend time with her Kevin Bailey founder of DreamFuel I think Manny will be there as well How many people will be at this dinner?

38:2050 and then Jen Agarchois got us to the guy's name is Ruslan something I don't know his last name but he's the head of AI for ServiceNow and he was the CMO of two companies previously So I'm really excited for that dinner. It's going to be a lot of fun. It's our 10-year anniversary. I think it's going to be a great event. It's going to be really galvanizing. So I can't wait for it. And we have Jeffrey Moore on the main stage. Jeffrey Moore. And we got Nick Maeda, who... Those are not in the same sentence. I'm sorry. I love Nick. But Nick's got a new business that he's going to announce on stage.

38:57I'm sure it'll be great. But Jeffrey Moore is like legend, legend, legend. This is the best dad ever. All right, guys. Pavilion. Join Pavilion. see us in New York City for September 28th through October 1st and otherwise we'll have more to talk to you about Pavilion in the coming months but it's going to be fun and if you want to meet me, Asit and AJ in person we'll be in New York in two weeks thanks everybody that they appended we're circling around something though that's like I think the really important piece to this which is you mentioned it Asit is around ecosystem and at the end of the day the CRM is the ecosystem.

39:34Well, I don't know if that's true. Like I, that's where I really disagree with that. That's the challenge we're talking about is that you have this hub and spoke model with Salesforce and distribution. And that's how these companies have really been able to get to scale inclusive of HubSpot, of course. And what you're, you're suggesting is that no, no, that's totally done. Someone's going to go horizontal. I think the ecosystem is dead. I totally disagree with that. I'll tell you why, H. Hardcore disagree with that. Because you have to like, if you take it from a first principles perspective, do we believe building things is easier?

40:10Okay, don't use fancy terms like first principles perspective to confuse me. Okay, let's just break it down. A lot. There's so many worse, there's so much worse nomenclature out there. Let me break it down. That's easy. Step one is, do you think things are easier to build? If shit's easy to build, then that was the number one reason that you would partner with somebody else because it was build, buy, or partner. I don't want to buy that thing. I don't have the money to buy it. I don't have the time and resources to build it. So I'm going to partner with that company. But now if building has become so much easier, you're going to build more of the things.

40:47So it leaves less room for partnerships. and hence the ecosystem opportunity becomes like if you're building a company and you want to be an ecosystem player you're on like some really like wishy-washy ground because you're like how do i know that what i'm trying to build isn't when i just get built in a weekend by this other company that i want to partner with and every month that goes by building things becomes even easier that's the one use case that's so powerful for ai so it's really hard to build that ecosystem company because that's you had to build it as an ecosystem you were like i am built to be a partner with all these other companies and i think building that company becomes really risky um partnership opportunities are going to go down more and more and it has to say that i would say that open ai and anthropic are ecosystem companies i mean that's that's the ecosystem what's that Who's the ecosystem?

41:47Well, right now, Granola's plugged into my cloud instance, and I've become quite dependent on it. What do you think an MCP is? What do you think an MCP is, Asit? It's an API. It's literally the whole point of it. I very strongly disagree with you, Asit. Very strongly. Let's let Keith talk about this. Keith's our guest. Well, it's funny. I think it's always more complicated. and I'll pick on call recording. This is a great example. And then I'll talk about sort of AI BDR. So with call recording, I think we have a pretty good out-of-the-box call recorder that works well for like a large SMB. It is not as good as Gong with all of the analytics and all of the sort of sales coaching that's built in.

42:37But we have an open platform. So some of our customers use our call recorder. Some use Gong. We partner with Gong, right? And, you know, and I think because of that... By the time you're a Series C company, Keith, do you envision Gond being something that a lot of your customers will need to use? Because you're right now, like you've been around for five and a half minutes and you have a good enough call recorder. You've literally not been around for more than a year, right? Like less than a year. Yeah, less than a year. And you already have a decent enough commission platform and a decent enough crawl recorder.

43:11And so like, if you play this out a year or two down the line, the world in which you would need a partnership with a drone or an outreach or any one of these companies, it just doesn't look likely that you will need them. You'd be able to build those things at a pretty good rate, which is really exciting. Like that's what you as a company should do. That's sort of his point about where you start, isn't it? I mean, isn't that your point, Keith? Because your advantage has to be, what do you view, you know, your moat as? Yeah, it's a good question. You know, the other point is we have a great granola integration.

43:49We're friends with the company. You know, I think building granola to granola's quality or building Gong to Gong's quality takes serious focus, right? Serious focus and R &D, I think, to sort of get to where it is now. I just want to sorry to interrupt but I just want to completely agree with you Keith like I think that that's people miss if you can build a product that's 10 times better for example with granola I know Asit hates it but I love it and when I get into a call now and somebody has one of those actual call recorders that like ding this call is being recorded I find it so tacky and so tasty I just because the smooth the beauty the elegance Asit from a taste an aesthetic perspective, that is why you should love Granola because the video conferencing interface is pure, you know, and it's not, you don't have like Keith's note taker app or whatever the frick, you know, in the side on the, on the zoom or the Google meet.

44:47And by the way, like now you've got a call coming up and Granola hovers and it does research on the web before your call. So it's like, Keith is the CEO and co-founder of Lightspeed. This is what you should know about their business. By the way, the last time you spoke to him, this is what he said. these are features. They're not hard to build per se. It's just about customer love, customer empathy. It is about how deep are you in with your customer so that you truly understand their workflow so you can build something that's magical, even though there's lots of people doing call recording. There's lots of people doing sequencing.

45:22Keith, what do you think about what I just said? Do you strongly agree with that? I strongly agree, which is to say, you know, for the next six months for me, where's my team focus? Where's my head? You know, it's really building out the business model, the business world model in the CRM. It's ingesting different signals. It's like building out more sales processes. It's like all of the role-based access control for different fields so the agent can go in. Like we're sort of all in on the system of record. And, you know, our goal is to be unbeatable at a bake-off on that note. And that doesn't really give us time or energy to go build the best granola.

46:03And I think that's sort of the nuance here, right? Salesforce released their AI BDR at Dreamforce. And does that mean the end of all of the AI BDR companies? Probably not. Are any of them doing well? I mean, it's not my bet. No, no, no. That place is dead. Okay, let me give you guys the talent argument. Firstly, AJ, you can only have a couple of things on Earth that are going to be like Claude or Codettes, which is like your main, it becomes your browser to the world. So not everyone can choose to be that. So you can't MCP, like every system of bracket will MCP into one of these platforms that becomes your gateway into the world, right?

46:49Sounds like an ecosystem to me. That sounds like an ecosystem. But the bar to be that, look at how high that bar is. These companies have had to spend, what,$100 billion on creating this high-level intelligence to earn the right to be that door. Not everyone can earn that right. But everyone is plugging into them and they're a part of the ecosystem, though, Asad. Yes, because they've done something so hard to earn that right. The bar is that high. How many companies will be able to match that bar? And how many doors to the world are we going to have? They don't have to because they built the ecosystem, which clearly shows the ecosystem is not dead.

47:22One ecosystem is good. We're not saying it's absolute that there's no ecosystem out there. What I'm saying is that the ecosystem strategy is like a business strategy. You're going to see it less and less. It's going to be less companies that are going to want to build themselves as an ecosystem player and less companies that will say my way of winning is the ecosystem. That's the bar. The other part of this that I think we are not taking into account enough is how easy distillation is becoming. Like a year or two down the line, you're going to take granola, you're going to give it to your cloud, you're going to say, here's a million dollars worth of compute, can you just distill this and create a complete replica version of this?

48:01And it's going to happen like this. It's literally, that is where the technology is going. So at that point, a lot of things that are standalone companies will become features. And the amount of product you're going to need to have on a company to be able to win a category is going to just be that much more. The volume is going to be... I will go with what Sam and Keith said, which is around the product. I think what you're saying, Asad, sounds very correct. It sounds correct, which is why... It's just very wrong. It would make me want to bet against it. like it's just so easy it's because it's it's like having pocket aces pre-flopping you like i'm gonna fold that i'm not saying to fold pocket aces but i don't i think i'm saying it's a smart sounding argument that is divorced from the reality of what it's like to work with somebody that wants a solution to their problem and needs and values something that's 10 times better that's what i for example right now you can just granola literally granola is not doing any you can distill it now it's you don't need a million dollars you don't you don't need a million dollars that compute it's a call recorder that runs on your hard drive instead of in the browser like it's fine you can copy it so why are they growing so quickly there's something else about it it's about the product team maybe it's about distribution it's virality maybe it's just a really good name we all love granola who doesn't it's healthy so at the end of the day building a company and building products is still really fucking hard.

49:33And that's what this is coming down to. And Keith, this is your second company. I mean, you spent five years at an AI first company that like you took to 41 million users and it was still really hard. So like tell us why he's completely wrong. Look, we buy a lot of software and I would say by any, any measure, we're a high growth company. But part of the reason we want to buy software instead of building it ourselves is because we want someone else to own supporting it, advancing it, making sure that it doesn't fall apart six months from now when I've got 30 reps and the last thing I want to do is troubleshoot the widgets that they need to get going.

50:15So I think that last mile matters. It's sort of funny if you were on Twitter for the past six months, every day someone said they were going to vibe code their own CRM and like, that's a wonderful idea. You should try it. You're not vibe quoting it for yourself. The argument is that if you chart this graph and you look at the last 10 years and you say for every stage of growth or for every million dollars of ARR, how much product did a company ship? What you will see is as we go from 10 years ago to now, it's increasing the volume of product at an average company ships per million dollars of revenue is dramatically higher.

50:58We can feel it. Like sometimes I go into a Series A, I think people have forgotten what Series A products used to feel like five years ago. Now I go into a Series A product and it feels like what a Series C product used to feel like. It really does. It's that much more advanced. It's that much more stable. It's really shocking to some extent. That's not going to stop. That's when I continue. I think the interesting thing is having looked, I haven't looked at your platform yet, Keith, but I want to and we'll be together in two weeks in New York and we can celebrate you. But I looked at a similar platform.

51:32The thing that strikes me about these go-to-market platforms, and that's why I still think there's an opportunity, is that everything's still the left menu. The interactions and the workflows do not feel naturally connected. It's like, oh, we built a sequencer. You can click that button. We built a call recorder. You can click that button. We built a forecasting tool. That for me feels disparate and it feels like there's still an opportunity to create some kind of united experience because right now, yeah, you can build anything and put it on the left side, you know, on the left menu and just like click on the thing that you want to do.

52:10But why are they in there in the first place besides the broad umbrella of go to market? And that's kind of what, I don't think that's exactly your thesis, Keith, but I do think that that's the benefit of depth versus breadth, which is the opportunity to go really, really deep on a specific workflow. Like anybody can create anything and just put it on the left menu or in Riverside, it's on the right menu. But how do you make these things integrated so that it feels magical and natural? I don't know if that makes sense. But yeah, that absolutely resonates because, you know, we we have competitors, obviously, and everyone's in a race to get, you know, the hundred features required to run a go to market team.

52:52And it turns out that that's really hard, you know, to get it to be performant, to get it to make sense, to get it to get those things to work together. and we've seen so many people that have like maybe just not wanted to put in the energy create this frankenstein app you know with a hundred features that in theory could prospect in theory could forecast in theory uh has a crm data model but is not a thing that a rep can use how is it raising money for this which is if you look back into the sass era obviously salesforce established SaaS as a thing. But since then, we haven't had all that many successes in go-to-market tech, right?

53:35Like HubSpot, ZoomInfo, kind of, you know, you could go into MarTech and find a bunch of stuff. But like, if you take the marketing piece out of it, there's not much. And if you talk to investors, they have, they, towards the end of the SaaS era, had PTSD from go-to-market tech. And now there seems to be this renewed enthusiasm for investing in go-to-market tech. We have Cassie at Primary, and she's got a big part of the fund's allocation is going towards go-to-market. You guys have raised a big round. There have been a number of other companies in the space. What was the experience like pitching investors on this?

54:18Were they ready? Had they been thinking about this? Was this a big thesis that they had? Did it take a lot of convincing? Because I can tell you a few years ago, it was a very sad experience for founders to try to raise funding for total market tech. AJ knows. It wasn't sad for me at the time, but I didn't raise it. I think this was true for any Series A right now. I think that there are these bimodal outcomes that investors adhere into. Like, for example, I don't see$10 million Series A's getting done anymore, or even$15 million Series A's. I think that the default is you don't get funded, or someone believes that there's like a$100 billion outcome here, so we're going to do the$50 million.

55:08And because of that, I think the companies that are getting funded all have some thesis around the system of record. They all have 0.1 % or 99.9 % percentile growth. And I think the investors know that it's going to be a little inefficient to get there because it's sort of this grab bag moment, right? Get every new company and grow with the best ones. And do we think that your team has the resolve to go to war with everyone else here and come out on top? So I think it leads to fewer companies get a lot more funding. Eiji, how is that compared to when you were raising in Zerp? not not that different in terms of i mean that there were 10 million and 15 million dollars series a's but it was still very similar in that looking at the outcome part of it i i think back to the thing that sam said though which is i actually want to draw back to because i'm curious keith if this came up in your fundraising strategy is just this idea of tying everything together which makes a lot of sense and i think that's actually a big part of what's gong's resurgence is um gong even their landing page you land there it's just it's like gong ai ask anything you want gong bridge just came out and so it makes this like very uh ubiquitous experience when you're going from account to account you just like you just can find all of your information really quickly it's not about just the forecasting or the calls or whatever it brings everything together and i'm curious keith if that one if when you're fundraising and go-to-market tech Like, is there like just like kind of this convergence and everything's consolidating and you have to pitch this$100 billion story?

56:54And then two, for your own platform inside Lightfield, what are you doing to tie all of those things together? Yeah, yeah. It's a good question. I think for Lightfield, the company, to make sense, you know, for an investor to write a$50 million check, but they sort of have to believe that it can get to, you know, that$100 billion outcome without ripping, you know, everyone off of Salesforce, right? You know, because I have a lot of marketing that says, you know, moving systems of record is easier now than it ever was. And that's true, but it's still hard and it's still annoying, right? So how do you get to a light field can have a big outcome without moving Costco off of Salesforce, you know?

57:39And I think it really is that you get much larger deals. Is Costco on Salesforce, by the way? Is that true? They're actually, I don't know. I imagine they are. They're definitely not on HubSpot, dude. They probably are. They probably are. They probably are. That's hilarious. I assume it's just like SAP at that level. All those companies are just SAP. That's my thought. I wonder if they're recording how many$1.50 hot dogs I've purchased in my lifetime. That number. Someone owns the Keith account. Got a caller in less than a million. Engage him. But, you know, to not have to rip and replace that size company, you have to believe the deal sizes will get really large in SMB and mid-market.

58:26And, like, how do the deal sizes get really large in SMB and mid-market? It's agentic work and consolidation, right? So a big thing that we had to show, you know, Andreessen, who let that around, was, like, you know, look at our deal size for a company that's 200 people or 300 people or 400 people. This actually looks like a deal size for Salesforce with like, you know, 5 ,000 people, you know, because of all of the work that we're doing. You know, so I think that was sort of the economic thesis. And then, of course, the product has to fit that, right? Because, you know, you'll only pay consumption pricing if I do great agentic work for you and you're happy with my sequencer, you're happy with my forecasting.

59:09How have you talked a little bit about pricing? Because it's a topic du jour. You know, it comes up in every conversation. And one school of, I mean, literally, first outcome-based or consumption pricing is impossible. CFOs will never approve it. Then it's, you know, we've got credits, but they don't know what they're buying. How do you think about passing through, you know, the compute and inference costs to the customer so that they can, and to Asit's point, confirming that Andreessen didn't care at all about margins. Did they ask you about your margins? But the bigger question is, how do you think about pricing?

59:49Yeah, yeah. They definitely asked for margins. I would say that we're not in a margin-optimizing state of the company right now. Thank you for your honesty. Yeah, we're trying to take the market, and my head of finance glares at me every day when we close deals. But there's a pop to it. You've got to play the game on the field, right? Exactly. On pricing, it's sort of funny. I tried both extremes. I think you learn a lot by trying extremes. We tried the sort of seats and platform fee, and that was the easiest to get through, you know, procurement. The problem was sort of the people at the edges.

1:00:29You had sort of folks that were using 10%, and then you and folks that were using 100 ,000%. And because of that, we tried moving to pure consumption. And then it was funny. It's like going to a store where everything's really extensive that our customers didn't touch anything. They were like, wait a minute, there's like a meter on every button that I pressed. And eventually where we landed was that You know, our seats and platform fee have to cover sort of core AICRM, you know, call recorder, transcription, filling out fields, tasks, you know, data model. That stuff has to be predictable. But there are things our system does where you can see the ROI on it and you're happy to pay consumption.

1:01:24The obvious ones are anything related to pipe gens. And, you know, it's pretty clear if you build lists and send LinkedIn messages that there's some ROI for you. Workflow automations, you know, the sort of AI ones where you're doing really complicated lead scoring. There's precedent for that. And then the last piece, which was sort of maybe the most interesting to me, is that most revenue leaders, CEOs, have an uncapped budget for business intelligence. You know, which is to say that if you want to run GPT-6 on your business world model to scenario plan, you know, what are we going to build next quarter?

1:02:08How should we change our sales process? You know, what deals are at risk? You will do that unmetered. I haven't met a single CEO that's willing to do that on the cheaper models. You know, they want to be token maxing and like squeeze out that like, you know, 0.1 % of alpha to sort of find truth. That makes sense to me, actually, because that's if you can push token consumption to the CEO, that is a person that will overrule the CFO and say, this is important to me. I need the answers to these questions. And if you can embed it into the product, then I can see I can see that being effective. Yeah, and I probably burned like a few hundred dollars on tokens trying to figure out whether we should do POCs pre-contract or post.

1:03:00And I ran like a hundred different versions of this and I think got to a very reasonable answer. And it saved me time. What was the answer? Yeah, where did you land? so below a certain deal size you know for us it's around 25k it's part of the contract you have to sign to to enable and then beyond that with good qualification we absolutely should assuming that i have enough implementation capacity to do it well uh what is your average contract size um right now it's sort of in the the 25 to 30k um with with some outliers and then yeah. And then we have some big ones where we're still figuring it out.

1:03:43You know, like I'm still figuring out how to do implementation and deployments and all of that. Keith, it's been great having you on the show. We really appreciate you coming, especially on short notice. And we're all three of us are going to be at GTM in New York in two weeks, where you will also be in your team. So we're really excited to see you in person. It's going to be a great event and we're really excited that you're partnering with pavilion and that you're sponsoring the show so thank you for that and we'd love to we'd love to take you know keep in touch and make sure that the product continues to grow we're really excited for your progress if folks what who's the icp you know we've got a lot of people listening if folks want to try lightfield who who should they i self-identify as and what's the best way to get in touch with somebody at the company so that they can try the product out?

1:04:35Yeah, I would say if your company is starting to grow or growth is your number one priority, we're a really good fit for us. So I would say if you're sort of a series B company or you're realizing that HubSpot isn't sort of, you know, enabling the work and forecasting that you need, we do really great work in that segment. I would say like a commercial team of 20 to 100 is sort of our sweet spot right now. And if you're interested, send me a personal email, keith at lightfield.app, or you could go to our website, lightfield.app, and one of our reps will respond to you in a couple of hours, because that's the rest of the length.

1:05:26A couple of hours. Okay, that's good. Within one hour would be ideal, but we'll talk about that in New York. Sam's not impressed with your couple of hours. A couple of hours? Why not three minutes? I've got some hours of customers to deal with. Well, Keith, thanks so much for being on the show. Folks, if you're listening, give us five stars wherever you rate your podcasts. Follow us and subscribe on YouTube. That's where we're really trying to build an audience. and we'll see you in two weeks in person, Keith.

1:06:01If you are looking for more Topline, check out the Topline newsletter at toplinemedia.substack.com. That's toplinemedia.substack.com. Or if you're a video person, check us out on YouTube. Topline-media is what you want there. Have a delightful day, everybody.