VC Trends & Predictions 2026

Julia Nimchinski:
Thank you so much. And we are transitioning for our next panel, BC Trends and Predictions 2026. Welcome to the show, welcome back, Kathleen Estrich. partner at PowerVC, and… An all-star panel here. Kathleen, how have you been? Long time no seat.
Kathleen Estreich:
I know, I’m good! How are you doing?
Julia Nimchinski:
Excited for this. I just have one question, and the stage is yours. What’s your top GTM prediction for 2026?
Kathleen Estreich:
Say that one more time?
Julia Nimchinski:
What’s your top prediction for AI and GTM for 2023?
Kathleen Estreich:
Oh. That’s a great question, and the topic of this panel, but I think things are continuing to get… more and more competitive, and I think this year was sort of, a lot of the hype, maybe not living up to it, but I think the technology’s moving very fast, and I think a lot of the promises of 2025 will actually become the realities of 2026, so… I think even in the prep for this panel, I was kind of comparing a bunch of the different, you know, co-pilots to help me, and it was just interesting to see how much deeper they’re going now than they were even, you know, 3 months ago, when you would try and, you know, do the same prompts. So, I think next year is kind of the year of enterprises actually adopting things because things actually work.
Julia Nimchinski:
Fingers crossed. Take it away.
Kathleen Estreich:
Alright, awesome. Well, thank you, Julia, for having me, and thank, thank you to the other panelists who are here. I’m Kathleen, I’m one of the partners here at PAIR. We’re a pre-seed seed fund, specializing, kind of, early stage on the pre-seed side, typically leading or co-leading rounds, and on the seed side, same thing. So, we do everything generalist fund. I focus a lot of my time on, on B2B, kind of, application layer. I’m really excited about the other panelists we have here today, so thank you all for joining. So I thought it’d be great to have you all give a very quick intro, and then we can dive in, so maybe, you know, your name, your firm, the stage that you invest, and then we can go from there. So maybe, Matt, you can kick us off.
Matt Hersh:
Hi there, it’s Matt from FICA Ventures. We’re, we’re also a seed in Series A Fund. We invest checks up to $5 million, In promising startups, we primarily focus on B2B. I lead our FinTech investment practice, and yeah, we’ve been around for about 8 years, and about $550 million in AUM.
Kathleen Estreich:
Jenny?
Unknown:
Jenny Fielding I Everywhere Ventures: Hey, great to see everyone, thanks for having me. Jenny Fielding, I run a fund called Everywhere Ventures, as the name suggests. We invest globally. We invest primarily across three verticals, money, health, and work. Jenny Fielding I Everywhere Ventures: And we’d go really, really early, so, you know, average round size, actually just crunched these numbers, was only, like, 1.1 million, so pretty small. We write small collaborative checks with folks like you. Jenny Fielding I Everywhere Ventures: Most of you on here.
Kathleen Estreich:
Cool. Kirby?
Kirby Winfield:
Hey, I’m Kirby Winfield, I run a fund called Ascend. 1 in 4 AI software engineers in the United States lives in Seattle. And I’m from Seattle, and I ran startups here for 20 years, and I’ve been running a fund for 6 years, investing in those AI founders. So we’re enterprise AI and deep tech, 500K check, inception stage.
Kathleen Estreich:
Cool. Twin Peaks, you know.
Sachin Patel:
Alright.
Kathleen Estreich:
Hey, it’s so…
Sachin Patel:
Sachin, the conference room here is making the background and the renaming difficult, so apologies for that to the group. Sacha and partner at Lightspeed Venture Partners, based here in San Francisco and Menlo Park. Lightspeed’s a multi-stage firm, sort of investing as early as pre-seed or seed, and as late as pre-IPO, which means check sizes anywhere from a few thousand dollars to, upwards of a billion at times. We invest sort of across the stack, consumer, fintech, now AI, infrastructure, application layer, model layer, etc. I spend all of my time investing at the growth stage, which for us means, sort of, Series B to pre-IPO, AI apps, and consumer fintech are where I spend, all my time.
Kathleen Estreich:
Cool. Bring it home, Praveen!
Praveen Akkiraju:
Yeah, great to be here, With Inside Partners, we are a multi-stage, investment firm. Everything from Series A’s through late-stage growth. I specifically focus on our early-stage investments, so Series A, Series B, though we’ve done a few Series C’s as well. Primarily, I have an operating background, so, I was an engineer at a startup. So, primarily focus on the infrastructure side of things, and, things like dev tools, cybersecurity, automation. Obviously everything to do with AI these days, but, yeah, Insights, been around for 30 years. We, our latest fund, Fund 13, is about $12.5 billion, so we have a pretty broad, check-writing capability, and typically we try to. Find companies, partner with them early, all the way through the journey.
Kathleen Estreich:
Awesome. Well, I’m excited to have all of you here. I actually hosted a lunch yesterday with, about 15 or 20 emerging managers, and talked about a similar topic around, kind of, trends for 2026. I think, as we look ahead, it’s always good to sort of think about the year that we just experienced, which I think was a pretty dynamic one in the world of venture and startups. So, I think one question that I wanted to start off with was, kind of. underwriting and how you think about defensibility, and particularly at the earliest stages. So I know, Jenny, you and Kirby and I spend a lot of our time with, like, very, very early teams, often, you know. some sense of an idea, but sometimes not even that. And would love to kind of kick off with you, Jenny, of, you know, how do you think about, kind of, moats when you’re making these bets on teams? What’s defensible? How do you think about that when you’re investing at the earliest stages, given, you know, there’s, like, 50 companies in every category, at least.
Unknown:
Jenny Fielding I Everywhere Ventures: I was gonna say, only 50, come on!
Kathleen Estreich:
Yeah, I know, yeah, in this one very narrow niche, but yes, I think there’s, like, many, many companies out there in every category, and then there’s a lot of, you know, it’s never been easier to build, so, you know, how do you kind of think about that when you’re underwriting these companies as the first check-in?
Unknown:
Jenny Fielding I Everywhere Ventures: Yeah, so this is obviously what keeps us up at night, defensibility, and I guess the way that I like to think about it, especially, you know, at the application layer, which is the only place that we’re investing, is this idea of, like, you know, a niche to a network effect, right? And so, if you can kind of zone in on a very specific niche, so, you know, maybe you’re doing accounting for real estate. Jenny Fielding I Everywhere Ventures: commercial real estate, right? Something that’s so specific, you can really build, like, incredible loyalty, brand, you know, brand awareness. It’s, you know, pretty easy to acquire customers, and, like, I would argue that it’s too small a niche, really, for the big guys to really care about or go after. So I think at the early stages, as you’re getting kind of your footing, not just thinking about product, but kind of thinking about, like, what that Jenny Fielding I Everywhere Ventures: is in AI, like, actually makes a lot of sense, right? And so, if you have a high switching cost, if you have a really engaged community that can then turn into that flywheel of a network effect, because the community is built, because, you know, there’s similarities in your customers, because they’re sharing the product, that has a lot of potential. Jenny Fielding I Everywhere Ventures: We don’t really know, I think is the Jenny Fielding I Everywhere Ventures: This is, like, my hypothesis on, you know, early-stage AI investing, but it kind of is too early to tell. But that’s, like, the questions that we ask founders at the early stage, and what we try to see. We haven’t really seen it play out, is the… is the honest truth. So, I’m curious what, what Kirby’s gonna say on this side.
Kathleen Estreich:
Yeah, Kirby, what are your thoughts?
Kirby Winfield:
I don’t know either. But, But I think, you know, I love the question around moat, because I think there is… kind of a new mode emerging for, for some of these companies, and we’ve made, I don’t know, 40 AI app layer investments, probably since 22. And the ones that you see really stick Kind of the question that we try to ask is, can you become a business unit within the company? that you’re working with, right? Within the customer. Can you do that within some 3-5 year period? Because it’s hard to… if you have a very high-performing business unit, you think about that as a, as an enterprise, that’s not something you’re gonna let go.
Kathleen Estreich:
Right? So there’s a lot of stickiness.
Kirby Winfield:
to that. And I think what we start with is, you know. Can you become a really high-performing employee? And then, you know, can you kind of grow out from there? And it’s hard, like Jenny said, like, we’re doing this, you know, it’s… we literally… the last investment we made, the guy just quit Microsoft, like, today. And we committed to the investment a month ago. So, it is very, very early, but what we’ve started to see is, kind of, if you start with getting… if you start with getting embedded in workflows. From day one. And, you know, there’s some… there’s some human in the loop, and it’s not… it’s non-critical processes. Maybe in year two, you can now start to be generating data that informs those workflows, and you get better and better at those workflows, and you get more trusted by the customer. And maybe in year three, you start to be able to provide some decisioning. And so I think that’s how we think about this… this idea of, like, a business unit moat. But, yeah, again, still watching it play out.
Kathleen Estreich:
For those of you on the growth side, like, how do you think about that mo… because you’re investing, like, you know, usually where there’s a little bit more signal, traction, customers, how do you think about, kind of, underwriting the… you know, and a lot of companies are going from You know, early stage to then, you know, the growth stage very, very quickly, and, you know, raising that… C, Series A, Series B, Series C, the growth capital is kind of, you know, kingmaking in a lot of ways very early. So how are you thinking about that in terms of what you’re looking for in companies around defensibility, and sustainably, you know, defending that as the companies grow at the stage that you invest? So maybe, Sachin, do you want to start? With that?
Sachin Patel:
Yeah, I’m happy to. I might give the same answer, which is to say, like, it’s almost too early to tell. Defensibility is evolving. You know, it’s funny, like, the patterns of what we all sort of looked at in the software context have really changed. You would normally look at things like retention and expansion, but The reality is enterprises are only just starting to adopt AI, so we don’t even have, like, a year of ARR data points to fall back on. And then it’s sort of the question of defensibility vis-a-vis startup competition, but also sort of the model layer companies as well, and then you can get into a bait of. what does defensibility look like if you’re a horizontal player? What does it look like if you’re a vertical player? On the vertical side, the things that we really look for are what is the fastest insertion point you can get into for your customers or for an industry? That’s typically shown in, like. very fast time to value, or ROI, so, you know, there’s a lot of ROI and voice right now that we’ve sort of spent time thinking about as a firm. So if you can insert yourself with, like, a voice application, and then build workflows, to Kirby’s point, around that, that’s sort of one way to look at defensibility and stickiness over time. But yeah, I mean, this is something we think about all the time, and I’d say it’s still early, and there aren’t a ton of, like, the traditional signals that we normally sort of fall back on.
Kathleen Estreich:
Praveen, anything to add from where.
Praveen Akkiraju:
Yeah, I mean, I think I broadly agree with, you know, the comments made, right? It is clearly early. You know, in some ways, as you mentioned, these companies are growing at a pace that’s much, much faster than what we’ve historically seen. with typical SaaS. And so, you know, you have companies that get into, you know, 10 million plus in ARR in a year, right, from launch. And, you know, I think, you know, one of the things that we do care about, again, I come from an operating background, I used to, you know, we did a startup, and customers, for me, is always sort of the North Star, so we tend to spend a lot of time with customers, both before making an investment in general. So, spend a lot of time with CIOs, you know, VPs of application, VP of engineering and such, like, to really understand, sort of, how they’re thinking through, you know. their tool stack, right, their, sort of, applications, and evolving them. And it’s, so… it’s a triangulation thing, right? We try to get the customer’s view, you know, we try to figure out, like. is there, to Sachin’s point, is that, how fast is the value realization happening, and is that sort of sticky, in the sense that, you know, there’s a certain category of companies that’s like, you know, works for the first few runs, and then afterwards, it’s sort of, you need a lot of a lot of, a lot of help to keep that thing going. I mean, most agents today are not quite good at long horizon tasks, so you need a lot of scaffolding, a lot of, you know, FDE and such like. So, trying to understand exactly what does it take for this particular product to be successful in delivering consistent value for the customer. You know, and I think the third piece, and this is something that we’ve all sort of come across, is this notion of the technology stack is evolving so rapidly. I think, you know, like, Gemini was, like, on the top of the, you know, software rankings for a week before Opus came out, so the pace of… Innovation is so rapid, and the infrastructure’s evolving so rapidly, that, we look for teams that are able to understand You know, how this… the underlying foundation is zigging and zagging, and can they keep up with that, right? And that’s a big part of trying to figure… printing out this picture. There’s no perfect answer, I think, as everybody said.
Kathleen Estreich:
Yeah, it’s interesting. Kind of switching gears, when you think about, it’s sort of… yeah, when you think about… Jenny mentioned, you know, looking for kind of that wedge use case that, you know, they can nail, and then, you know, the workflows were, you know, were mentioned, so you land somewhere and expand somewhere. I think, like, 2 years ago, everything was a GPT wrapper on top of something, then that went out of style, you know, for a year, and now it’s kind of back, but being very purpose-built for very specific verticals. the big foundation companies are doing a lot on the horizontal side and, you know, kind of owning some of that stuff. I think Google in the last couple weeks is sort of coming from behind, well, not really behind, but, you know, coming up and making a ton of improvements. So, how do you think about, kind of, feature, like, nailing that one early wedge use case, you know, versus building a platform, versus building the the workflows, like, what is sort of your firm’s… stance on, kind of, feature versus company, and, like, maybe it’s at different stages, you’re looking for different things, so maybe, Matt, you can, answer that one first. Like, how do you think about, like, feature versus company versus platform versus API, and sort of the things that you guys are specifically excited about, and then things that you won’t touch, You know, based on what’s happened the last few years.
Matt Hersh:
Yeah, I mean, look, like, all of us on this call are trying to make investments in platform, right? I mean, I think the features are the ones that we worry about, right? The feature-only companies, right? That’s where you come up with a… you find yourself in a situation where there’s 50 companies building all these features, you know, the platform is the critical piece, and platform is sort of a vague term, but I think it… I think the definition is actually how Kirby chose to answer it about being able to penetrate a… a customer… type or a department in a company, right? Because that’s when you become… you sort of are able to lay the bedrock, for that platform. Also, platforms typically require not only the software services and data that your company may provide, but… but oftentimes the mode is created when you’re able to integrate with your customers’ data sets, right? Maybe a simple example is just being able to push and pull data from from their CRM, right? It’s very sticky, it’s very tough to unwind that if it’s a platform play. I think the riskiest model is the API model, right? APIs are extensible, they’re flexible, they’re beautiful. They sort of allow highly customized experiences, but you can flip them off with a common in the code, right? And so… you know, we always look to invest in those businesses that are truly platform, and that, you know, even if we don’t see it in the early days, like, we have to believe that at some point that this can, you know, be a GUI, that is used or integrated into the existing systems at the customer.
Kathleen Estreich:
Yeah, sometimes I think about it as, like, you know, it’s like Act 1, Act 2, Act 3 of a company. Like, Act 1, it’s like, nail this, like, very narrow use case, get a bunch of people who love you, then act two, it’s, like, build from that. Act 3 is, like, you know, be the end-to-end, but it’s hard to, you know. go from Act 1 to Act 3, and then, you know, anyways, so, Kirby, you were smiling, what’s your response?
Kirby Winfield:
Oh, yeah, no, I mean, it… a lot of this is just… you know, same as it ever was, right? It’s… it’s… what we always tell our founders, like, customers don’t care about your AI. They care about solving their problems, and… and it’s about trust, and if you can solve… I mean, Satya and I are in an investment together, actually, In a, in a vertical, AI Agent 4. financial advisors, and, like, they came in and solved one problem, which was secure communications.
Kathleen Estreich:
And…
Kirby Winfield:
Now, because they did that so well, their customers are asking them to solve other problems. And, like, so what we try to do is… is bet on founders who are gonna follow that customer need, and, well, first build the trust, and, like, nail that initial use case, because now, more than ever, there’s, like, there’s really no excuse to… I don’t forget who… somebody posted on LinkedIn. recently an investor I really like, and I wish I could remember who it was, but that there’s, like, maybe it was Steve Vasalo from Foundation, but there’s no more excuse for, like, a minimum viable product. and I forget what he coined, it was something like Minimum Fantastic Product or whatever, but, like, the point is, like, we have to have confidence that a founder’s gonna… that their first swing is gonna be a home run. Even if it’s in, like, a small ballpark. to, like, a power alley, and not… you know, it’s gotta be… it’s gotta be there, because that’s what buys you the trust to be able to then go out and find those other problems and build for them, and eventually then when you solve those problems, you get the trust of the customer, where they don’t ask you to solve specific problems anymore, they wait for you to tell them what problems you need to have solved, and then your platform. I don’t know.
Kathleen Estreich:
Machin, anything to add?
Sachin Patel:
No, I think that was well said on both. We think in acts all the time. It’s particularly important at the growth stage, because You were paying big prices in today’s market, so you have to believe in, sort of, the multi-act story of these companies.
Kathleen Estreich:
Yeah, that kind of brings me to… I think there’s a set of companies that are kind of, you know, the 5 to 10 years old, not AI native. Some of them have made super high, hard pivots where they’re, like, AI, all in. Others have been more hedgy. I’m curious, kind of, how those of you who’ve kind of been in venture with maybe more mature portfolios, like, how are you advising these companies? Because it’s kind of… they have, like, 6 to 12 months, in my mind. until they’re… they’re irrelevant. So, Praveen, you’re, nodding your head. So, I mean, Insight has a great, you know, big, great portfolio. Like, how do you think about that in terms of, you know, these companies that… have been started, you know, kind of pre-Gen AI.
Praveen Akkiraju:
Yeah, no, it’s, that’s the, I guess, the trillion dollar question, right? because all of us have these, you know, portfolios, particularly for a growth stage investor, of some phenomenal SaaS companies, right, that were, you know, maybe 5 years old, 6 years old, that, that we are now trying to… that are now trying to navigate this sort of shifting foundation that we were talking about, with AI. And, you know, I think, in some ways, it always comes back to the North Star, right, which is, you know, customer value, and, you know, you heard all of, you know, everybody say the same thing, right? It’s like. How you deliver customer value? And, you know, yeah, I mean, AI is a… is a game changer. It fundamentally transforms a lot of, you know, a lot of aspects of how you build a product, from the user interface down to. you know, the kind of context you can bring to it, etc. But ultimately, the customer cares about solving their problem. So, these companies that have been doing a phenomenal job, you know, as SaaS companies, I don’t think necessarily have an existential threat? Because it comes down to, are they smart enough to understand How do they leverage this new tool called AI to, you know, accelerate their customer value, right? And in some cases, you’re right. I mean, in some cases, it is a hard pivot. You have to basically recognize that, okay, like, you know, I’m building a lot of legacy software, you know, I need to fundamentally change the value proposition, it has to be agentic. In other cases, you know, it is about basically enabling a different you know, a different set of value for the customer by adding AI as an interactivity layer, or by making it more dynamic by bringing in, you know, different tools and such like. So, I think the smart founders are trying to figure out what does AI mean in their context. I think the more interesting question, beyond… so there’s the product aspect of things, and, you know, hopefully we bet on founders that are close enough to the customer and close enough to technology that they can intersect it. The more interesting one, I think, is the business cases, the business models are evolving, right? AI. And I think that, in some sense, is more interesting, because now software has a variable cost, right? You know, at scale, obviously, you know, the model pricing is still sort of on this downward sloping curve, but So there is a fundamental set of primitives that are evolving on the business model side that you need to think about and make sure, like, how do you think about gross margins in the era where you need a lot of forward-deployed engineers, right? How do you think about, sort of, retention, where maybe there’s a little less loyalty? So, we used to have these benchmarks in, you know, standard, like, NDR, GDR, whatever. Right? Magic number, etc. So, smaller companies, smaller teams, building more value. So, I think there’s two aspects. We think about it both on the product side, as well as on the business model side, and try to kind of, you know, make sure we’re tracking those.
Kathleen Estreich:
Does anyone have a portfolio company that’s done, like, yeah, like, an example that you think of a company that’s done this very well?
Unknown:
Jenny Fielding I Everywhere Ventures: Well, I mean, I have, like, an anecdote, which is, like. Jenny Fielding I Everywhere Ventures: actually, if you went back to a lot of the SaaS companies that you invested in in 2017, 18, 19, and you looked at their decks, their decks actually, many of them would be AI first. And I did this exercise, I went back to some of our companies that were past Series B SaaS, Jenny Fielding I Everywhere Ventures: They actually had a founder, or, you know, a very important person on the team who was looking at that, and as the company progressed, and they started taking capital from, you know, some of the big folks, including people on this call. Jenny Fielding I Everywhere Ventures: the priorities shifted, right? And some of those people were sidelined, so I’m not gonna name the company, but we actually have two companies in our portfolio where the technical founder was a little bit sidelined over the last years, because Jenny Fielding I Everywhere Ventures: they were, like, just loving building cool stuff, including, like, very AI-native, but, like, the technology wasn’t quite up to there, and the cash wasn’t allocated to that. Jenny Fielding I Everywhere Ventures: all of a sudden, those founders have, like, come back and been completely elevated, and the companies are really orchestrating around them. So I think it’s a great exercise to, like, go back, look at, which of these SaaS companies always had the plan to, I mean, obviously not become AI-native, but that was, like, always the goal, and figure out, like, when the direction shifted, and I think you’ll see at least 50% of them had that plan, and it kind of got.
Kathleen Estreich:
sidelines. So I’d love to hear some of the later stage investors just talk about kind of what they’ve seen in their portfolio from that test, because we did that, and we had, like, two companies that are kind of transitioning now, and the kind of original founders are really coming back in to lead that charge.
Unknown:
Jenny Fielding I Everywhere Ventures: Successful.
Kathleen Estreich:
When did they get away from it? Like, was there a consistent time, like, based on what you observed? Or was it sort of gradual? I’m just curious.
Unknown:
Jenny Fielding I Everywhere Ventures: Well, I mean, once they hit Series A, and, like, $20 million came in, and, like, you know, the priorities of the company kind of shifted to just, like, all about sales and piling into, like, what was working, as opposed to developing new technologies and innovation, which was the goal at the beginning, because that’s why those founders joined. They didn’t want to just build, like, boring SaaS companies, like, they really wanted to kind of create something Jenny Fielding I Everywhere Ventures: that was really more of what we’re seeing now, it’s just the technology was kind of hard. So, I think it’s kind of exciting to, like, get those people back in the forefront of those companies. Jenny Fielding I Everywhere Ventures: Some of them never had it, I guess, but a bunch of them really did.
Kathleen Estreich:
Such it.
Sachin Patel:
What have you got?
Kathleen Estreich:
What have you seen? I think that was directly at the growth investors. What did you do to mess these companies up?
Sachin Patel:
Yeah, I mean, I can give two… I can give two, late-stage examples, and these aren’t, like. Lightspeed-owned views. The two companies are sort of, like, very well known, and I don’t want to say consensus, but well talked about. You know, we’ve invested in RAMP very recently, and then we’re also investors in Databricks, and Those are two companies that I think everyone would agree started pre-ChatGPT, pre-AI, etc. You know, and sort of the RAMP example, when we led this most recent round, what got us excited was goes a little bit back to what Praveen said on team, right? Like, do you have the people in the building, who can see around corners take, sort of, the strategic turf that they’ve built and established over the last 3 to 4 years, and then apply AI to it, both in how they work internally and how they deliver products to their customers? And that was… that was really clear for RAMP, right? Like, Act 1 of the company was… was payments and spend management, and Act 2 was bill pay, Act 3 was actually just a software product. And now they’ve sort of had the talent and team to take all of those inputs around payment data and software and layer on Agentic and AI products on top of it, so that’s, like, a well-known and well-talked-about example that’s sort of… existed pre and post-AI and gone through the journey of multiple valuations. Databricks, very similarly, was, like, classic software, and, you know, Ali has made some really smart acquisitions and built some really great products to position them for, like, AI infrastructure in the future, so… Those are easy examples to point to because of the success that they’ve had, but there… I think there are lessons from those companies that we sort of talk about with our portfolio founders as well.
Kathleen Estreich:
We had in a couple months ago in one of our, kind of, quarterly portfolio reviews, sort of, like, who are those companies that we need to go have that conversation of, like, hey, if you don’t go all in on, you know, making this more agentic and more, kind of, AI-forward, you’re gonna be disrupted, and, you know. I had to have some of those conversations, it was interesting.
Sachin Patel:
though, right? Like, Privine said, I don’t know that we’ve seen categories be totally eviscerated by AI yet. I mean, you could argue some education companies without naming them, you could argue, you know, people thought Google and search was gonna be eviscerated by perplexity, right? And, like, that hasn’t… happened yet. So I think it’s just about staying nimble and adapting more than Is my company or category gonna disappear?
Kirby Winfield:
I mean, I think one thing, I’ll just say one thing we’ve seen is we have… we have a legacy investment, I mean, like, Fund One, it’s been around for, like, 8 years, B2B marketplace, and it’s good. It’s fine. But, what’s really interesting is they… started building their own AI tooling to, like. More quickly process, and more efficiently process, and assign jobs to folks within the network, and… Like, they started turning it inside out, and… selling it to their big job demand customers, who are in a very old-stage kind of insurance-adjacent industry. And it’s the first, you know, it’s the first time anyone’s bothered to pitch these guys, and so they’re starting to get traction on the AI side, and so now it’s like, okay. good, because, you know, the multiple on your marketplace business sucks, but also, like, that keeps the lights on, and now you’re riding two horses, and investors aren’t gonna like that. So it’s a really hard… even if you can get… some of that pivot done, like, it’s still… it’s still very hard to pull off at much earlier stage in the companies that…
Kathleen Estreich:
Handshake is now, you know, they were the recruiting platform for college kids, and now they… they’re leveraging, parlaying that into being, like, a Mercor competitor, which is pretty interesting, so…
Sachin Patel:
That’s right.

Kathleen Estreich:
Yeah. When you guys think about, kind of business models, so back to the point Praveen made, I think, like, the pricing and, you know, service as a software trend, around not just going… tapping into software budgets, but also, kind of, personnel and people budgets, Where are you seeing that work, The most, kind of, within your portfolio, and what are your predictions on, sort of, how pricing changes, kind of, in the next 12 months, in terms of a lot of these products kind of going to market, and feel free to kind of use real examples of companies that you’ve seen do this well, whether it’s in your portfolio or just generally. Anyone want to kick that off?
Praveen Akkiraju:
Yeah, I mean, I guess I could… I could take a crack at that. Yeah, I mean, you know, I think we, you know, as an industry, software industry, we’ve kind of, you know, come from the shrink wrap, the ELAs to subscription-based software, and, you know, if you read the… if you’re on X enough, you’ll… you know, the tradition is towards outcome-based pricing, right? I think what we’re saying is, the reality is, it’s, it’s a great aspirational goal to get to pure outcome-based pricing. It’s a hard. Thing to execute, because, both as a company or a startup delivering the outcome, or delivering the value, to be able to kind of capture that in a way that’s not seed-based or, you know, amount of database, amount of data processed, is, it’s, it’s… it’s a challenging thing to nail an outcome-based pricing model, unless, of course, you know, you have something very clear, like customer support or, you know, calls deflected and things like that, right? On the other side, actually, we hear from enterprises, right? The CFOs have been now gotten comfortable with subscription-based models. They can budget for it, you know, they can report on it, you know, and they know how to build the financials around. Now you have an outcome-based pricing model how do you actually create a budget, right, for… let’s say you have, like, 100 agents, and all of them are outcome-based pricing, and, you know, with varied number of users, and spikes and such, like. So, I think we’re, as an industry, broadly speaking. we’re still kind of working through what that ideal pricing model kind of looks like. Some of the examples that I’ve seen is, you have you kind of split the baby, right, in the sense that you have a platform-based pricing, so it’s like, hey, here’s my agent-based platform, so if you look at Workato, one of my companies, they charge, like, a fixed price for their iPaaS platform. It has all the integrations and the agent framework. But then they have an outcome-based pricing model for the agents that are built on top of the platform itself. Right? And so it’s a bit of an experimentation to kind of see how the customers adapt to this versus, you know, whether we’re capturing value. So I think we’re, you know, just like with the product itself, like, we’re going to iterate and learn where we land, but I’m not sort of fully bought into this notion that we’re all going to go to outcome-based pricing. Just because now it’s AI.
Kathleen Estreich:
Jenny, you guys have a pretty big portfolio. What are you seeing, kind of, across your portfolio, and what’s working?
Unknown:
Jenny Fielding I Everywhere Ventures: I’d say… at the early stage, too early to tell, but we haven’t seen, like, a huge shift in the later stage companies yet. I mean, we’re kind of waiting for it, I think people are prepping, but, quite frankly, we haven’t seen… we do a lot of, you know, turnkey SaaS, and we haven’t seen much innovation there yet, although we hear about it.
Kathleen Estreich:
Matt, you focus a lot on fintech. What’s happening, kind of, with pricing in the fintech world? Feels pretty dynamic.
Matt Hersh:
Yeah, it is. I mean, I don’t know if it’s… necessarily evolved as much. I mean, the fintech, you know, if you… if you think of, like, a pie, right? The disruption in fintech happens where you can try to figure out how to get a big slice of pie. And so, you know, that’s typically would boil down to, like, interest, or BIPS on interchange, or… I don’t know, like… cost types of variables, right? Like, what the default rate is, and things of that nature, if you’re, like, a lending business. Or, like, take rate, or the rev share if you’re a reseller of, like, an embedded fit product from a fintech perspective, so… I’d say those business models are pretty… codified. You know, the worry with those fintech business models is that they end up being a race to the bottom, right? That, you know, it’s a… your customers demand the economies of scale to be built into your pricing models, and as volume increases, pricing goes down. And oftentimes, if you’re… You know, built on top of a… call it a processor or a network, you know, there’s… it’s a catch-22, because your… your cost, you know, structure may be upside down based on, you know, what you charge. In those early days, so it takes a bit of a leap of faith from… The executive team and the board, where, like. we will sometimes, prioritize the quality of logo, and frankly, the quality of the contract. I mean, I’m a former salesperson in enterprise and in banks and fintech, and so I like to get in the weeds with our companies on their on the clauses and the terms that they agree to, right? Indemnification, what liability is, termination for convenience, right? Like, all of these are, like… and frankly, like, the marketing rights, like, can I talk about this customer, right? Can I put their logo on my website, right? There’s… there’s a lot more that goes into, like. pricing than just the numbers. And so we’ll sometimes sacrifice, like. Optimizing price, if we can, like, win some of those other, sort of, high-value Kind of variables, as part of a… customer contract negotiation, but, you know, I think those… those… the reason why FinTech has been hot, is because those, you know, it’s sticky. The services… you know, tend to be an annuity, and, like, a bank or a large fintech or vertical SaaS company struggles to rip out the fintech products they embed, and that’s sort of the beauty of the model from our perspective.

Kathleen Estreich:
With more and more, kind of, products coming online. you know, 50, 100 companies per category. I think go-to-market is mattering more and more, and kind of how you get unfair distribution, whether that’s, you know, through partnerships, ecosystem, channels that work, brand, founders who are famous, who can, like, you know, get that unfair distribution. When you think about, You know, your companies across, you know, bringing that product to market across the go-to-market stack. what are your predictions for, sort of, what… you know, this year, I think a lot of stuff it was like, outbound is dead, it’s not really dead, it’s just changing. So I think there’s a lot of headlines, and then there’s reality, and I’m just curious, sort of, you know, this is a predictions, looking ahead to 2026, like, what are you all excited about for your portfolios in terms of go-to-market, in terms of what’s working, what you’ve seen, you know. starting to be differentiated there. My thesis is, like, I think the best product doesn’t always win. You also have to win on go-to-market, and so I spent a lot of my time in the early days, like, testing, do these founders have good instincts around how to bring products to market? Because I think you need that in this pretty noisy world. But, yeah, curious to see, you know, based on from where you all sit, like, what are you seeing from The go-to-market stack and what’s working, and how you sort of test for that when you’re thinking about investing in companies, or doubling down on companies.
Matt Hersh:
Maybe I’ll go real quick, since, you know, I spend all… most of my time supporting our companies posting.
Kathleen Estreich:
husband.
Matt Hersh:
And have… did operate as a sales leader for many years. What’s happened with the velocity, equation that Praveen mentioned earlier is that The messaging is decaying at a fast… the sales messaging is decaying at a faster rate than the funnel is developing, the sales funnel is developing. Right? So, like, I think that the biggest challenge companies have is they just don’t update their materials, they don’t update their website, like, as the product evolves. And if, like, you’re a high-power… like, high operating engineering team, where you’re shipping new features, like RAMP, for example, on, like, a weekly basis, if your sales and marketing team isn’t right there with you, hand-in-hand and lockstep, you’re gonna… you’re gonna not be able to keep up from a go-to-market perspective, right? So… now, one might say, well. you know, you find yourself in a situation where you’re overselling, right? Because you’re talking about all these capabilities and, like, the future of what the product is going to look like at some point down the road. I think that can be done in a transparent way, and it can get customers excited, both new and existing customers from an upsell perspective, but, you know, I think it… the strengths of the go-to-market team have to be, like, almost like a translator, right? They have to be really good at translating what’s happening within the product and engineering org to what’s actually happening with their boots on the ground as they go visit customers. And being able to map, you know, the company’s capability set to Those customers and their workflows.
Sachin Patel:
Yeah, I agree with that, Matt. There’s, like, there’s three things that we talk about. One is, what Praveen mentioned earlier on the sort of ROI equation for the decision maker. So, what legacy budget are you taking, or what sort of productivity are you giving to your users, and how does that translate to pricing? Then there’s sort of, like. the implementation and integration component of go-to-market for deployed engineering is sort of the buzzy term right now, in the legal domain. It’s legal engineering, right? And it’s like, to get that ROI, you have to be, you know, pretty embedded into various workflows and data in the company, and then the third leg of go-to-market, at least for us, is sort of user sophistication and education. Like, even speaking personally, like, I don’t know that I fully use all the AI tools available to me to my advantage, so, you know, you think about people in… whether it’s legal or other sort of verticals, there’s a lot of just, like, user education required as well to extract value out of the tool, so I think the best companies Think about go-to-market on those… those levels.
Unknown:
Jenny Fielding I Everywhere Ventures: I mean, I’ll just add, I still think there’s a lot of hesitation with B2B founders at the early stage to spend any time on things like brand, and that’s something that we try to, like, shake out of them a little bit, because even with, like, no budget, you can create a really strong brand in a very unsexy. Jenny Fielding I Everywhere Ventures: category, by LinkedIn posts, by, you know, having some conferences that are kind of good for your users, your community. And we’ve just seen, like, it paying such dividends with companies really early stage. You just mentioned, Jenny Fielding I Everywhere Ventures: legal tech, we have a company called Udia, and, like, they’ve now raised, whatever, $100 million from General Catalyst, but, like, at the beginning, it was just, like, the founder getting on LinkedIn, just, like, talking about, you know, the way he saw the world, and we just encourage our other founders to, like, think about that, because it developed a very strong, sticky brand. Jenny Fielding I Everywhere Ventures: Around what his message was, and it really, literally cost nothing. So, I think there’s still a lot of, you know, oh yeah, you’re a consumer, of course you’re gonna invest in it, but brand at the early stage for B2B has a lot of power.
Kathleen Estreich:
Yeah, I agree with that, and it’s not… it’s not expensive, it’s just, like, also figuring out what makes you unique, and having something unique to say in the market, which goes to, like, why are you starting this company in the first place? And having something interesting to say, but yeah, I think the company… I think it feels like a big investment, but, when you actually put dollars behind it, it’s not actually that much money, it’s a lot of time, because there’s a lot of free channels out there, and the tooling for go-to-market is so much better than it was. You know, I was an operator for a long time on the marketing side, and man. What you can do now is just amazing, with a very, you know, very shoestring budget, it’s amazing what’s possible. Kirby, anything on your side?
Kirby Winfield:
Yeah, I mean, I guess, I guess it… Two things, one, to Sachin’s point, like, the thing that you asked what we’re excited about for 2026, and, like, what you think trends are, and I think… On his second point around embedded engineering, or forward-deployed engineering, or what have you, just, like, services, like, bespoke onboarding. I mean, I think enterprises are… have realized you can’t throw LLMs at stuff and expect it to work, and I think that they’re just realizing that, from as far as I can tell. And I think what that means is that they’re going to be a lot more thoughtful about who they engage with and how they engage with them, and so I think you know… Consultative sales. And patient… You know, the cycles are probably going to be longer than you want. But the dollar amounts that we’re seeing committed to these first engagements, and, you know, most of it’s software revenue, some service, but, you know, it’s $250, it’s $500, it’s $700. For, like, an for an app layer agent company. And that’s… I don’t know about Jenny, but, like, for me, with SaaS, I was used to, you know, I’m used to, like, those first contracts, 25K, 50K. And so, if you can land those with good logos and secure them and expand them. Yeah, you might have less of them because it takes so long, but that is going to be impressive, durable, sticky revenue, and you can start to now pencil your way to 100 million AR a lot faster. after a very slow, sometimes, trip to the first 2 million, if that makes sense. And then I think, yeah, just to…
Kathleen Estreich:
There was, I don’t know if you know Vili from Category, he posted something recently about forward deployed engineers, and it was, like, some skepticism, because I think it’s being used in the wrong ways sometimes, frankly, with founders, where they’re like, oh, we have nothing, but we’re doing the forward deployed engineer strategy, and you’re like, okay, what does that mean for you? It’s like, you’re basically a consultant right now, versus it actually… I think the way that Palantir did it, who, you know, made it, you know, somewhat palatable from a venture perspective, I think people are misusing it in a lot of ways.
Kirby Winfield:
Yeah. Yeah, yeah, no, I agree, I think…
Kathleen Estreich:
about it, like, how do you see it actually working, where you don’t.
Kirby Winfield:
Yeah.
Kathleen Estreich:
stuck in this services sort of rat race, where there’s not a product. Yeah, you can’t…
Kirby Winfield:
You can’t, you cannot conflate, Pilots with 4 deployed engineers, like. that’s not the same thing. So, where we see it work is when a company has spent a year and a half Like, building something. Where there’s a there there, and where they’ve had pilots.
Kathleen Estreich:
And…
Kirby Winfield:
just the realization, I think, for a lot of these companies is once you’ve done that. you’re not… you’re not… you’re still not ready to sell, like… you’re not a platform, right? So it goes back to our discussion earlier. Like, you’re just trying to get… two agents to deploy it at the enterprise. Are you trying to get… you automate… try to automate five workflows. And, like, it turns out, with all the security reliances, and all the decision-making that needs to take place at, like, a major pharmaceutical company, or, like, a top-five bank. Like, that… it’s 3 months just to get the spec done, even if you already have built the product.
Kathleen Estreich:
For that engagement.
Kirby Winfield:
And I think that’s when that investment makes sense, and that’s what we try to kind of guide folks to, is don’t… don’t, like, hide an unfinished product behind the… behind the, you know, window dressing of a forward-deployed engineering strategy, but… but when it… when it’s right, I think it’s right. And then… and then just last thing, I think to… to the… to Jenny’s point. And to your point, around how cheap it is to do go-to-market, you know, brand marketing now. We tell… I mean, we tell…
Kathleen Estreich:
founders all the time, and, like, literally 1… 1 in 10 maybe listens, maybe? Maybe.
Unknown:
Jenny Fielding I Everywhere Ventures: Thank you. You can… you can, you can just…
Kirby Winfield:
do, like, an 8-question survey, and you can send it to the person at your ICP company, or the ICP, You can send it to them, and make sure it’s the junior person who hasn’t been featured on any articles or LinkedIns or anything. Send it to them. They’ll fill it out, they’ll put it… they’ll send it back. You put it on your blog. You put a picture there. You write an intro. They post it on LinkedIn, all their friends like it, because people like to like their friends’ stuff. They’ll boost it. Do it… do that 200 times in a year, in your category, and you will have a brand. And it’s like, yeah, but… and, I mean, I don’t know. I… so I wholeheartedly agree.
Praveen Akkiraju:
Oh my god.
Kathleen Estreich:
investment in a website, it’s like, even if you’re doing outbound and your customers are enterprise, your website actually matters, because people are seeing you on LinkedIn and then going to your website, and I think making your website very clear, matters, even if no one’s buying your product on the website.
Kirby Winfield:
I agree. But I’m a recovering marketer, so…
Kathleen Estreich:
Me too.
Praveen Akkiraju:
I was going to add a comment on the forward deployed engineers, because this is something that… it’s kind of interesting to me, right? In some sense, we’ve always had forward-deployed engineers. They were called solutions engineers, or technical marketing engineers, or whatever, right?
Kathleen Estreich:
Talk about branding! Rebranded!
Praveen Akkiraju:
Right. You know, it’s just now more, cool, right? I think I saw, was it, I saw an article from one of the venture funds who basically said, the hottest job in the market right now, FTEs, right? But, you know, it also kind of tells us something, right? You know, it tells us that a lot of these AI-native products are, are having a hard time getting customers to value out of the box.
Kathleen Estreich:
Right?
Praveen Akkiraju:
And that’s because I think we promised something more than what we used to promise with SaaS. With SaaS, you know, it’s like, hey, here’s a piece of software, it’s running on this website, you know, you need your engineers, or your business analysts, or whoever it is to create the workflows to get the value out of SaaS. It was kind of baked into the assumption. And if you had Salesforce, you had a Salesforce implementation that came with Salesforce, right? Whereas, I think with AI agents, you know, we promise that, hey, these things do work for you, right? They’re supposed to automate away a lot of this sort of analyst tasks that are going on top. And to actually make that work, you need a bunch of integrations and things that Kobe was talking about, policies, governance, etc, all that stuff that needs to be bootstrapped before you can actually get the value. So, effectively. there is no way for this AI agent to deliver value until you get it set up, right? And that’s sort of now falling back to the company, so instead of.
Kathleen Estreich:
The customer hiring a consultant to do the implementation, that is now.
Praveen Akkiraju:
you know, on the… on the company to make that implementation work. So, this is what, again, talking about business models, right, you have to kind of rethink this, and I heard this, like, oh, AI is a services business. It’s like there’s a lot of these things being thrown around, which… which are, you know, it’s telling you something, you just gotta, like. First principles, trying to understand what’s going on.
Kathleen Estreich:
How do you underwrite that, though, as a growth investor? Like, how do you view revenue that’s, you know, the software versus services, you know, how are you thinking about that?
Praveen Akkiraju:
Yeah, I mean, I think for us at Insight, we’ve not had, like, a negative view of services-based revenue. You know, as a CEO of a company, you know, we were like… at one point, we were like, 21% of our revenue was services, and the board would beat up on me, saying, I gotta get this thing down, and eventually we were at 15%. That was the right balance, right? But I think it’s really… Because we’re early days, we haven’t quite seen what a forward-deployed engineer model means to the business model of the… of the… of that company, right? Because somewhere… is it… is it R&D cost? Is it like an engineer? Or is it actually, you know, COGS? How do you classify a forward-deployed engineer? And so I think we’re still kind of figuring this out. At this point, the focus, again, is on accelerating customer value, and so, you know. we… we… well, like, whatever it takes to make sure the customer realizes the value, and then if you find that as a pattern, you’re just going to have… have to hand-hold the customer, like, throughout their journey, then clearly there’s something broken in the way the product is built, right? So I think it’s, again. we’re kind of learning through this as, you know, everybody was saying, like, we’re so early in this, right? Some of this is just going to get, you know, worked out.
Kathleen Estreich:
Yeah. So, we have about 5 minutes left. I want to make sure that you all have a chance to give your, kind of, prediction of 2026. So, I’ve got a two-part question that everyone has to answer. The first is, what is one sector that you think will have the most unexpected, kind of, acceleration of AI, and what is one potentially long-standing category where it’s going to become uninvestable? And commoditized, basically. So, Who wants to kick us off?
Praveen Akkiraju:
I guess I can go, that way I have the advantage of saying the most obvious thing. But, you know, I think, everything is going to go multimodal. I think platforms that incorporate multimodality are going to see this tremendous acceleration. Today, you see it primarily in the customer success kind of use cases, but I do think that, directionally, if you look at the big steps that we expect the foundation models to take. It’s in the area of multimodality, and I think platforms that incorporate that will see significant acceleration. I think we’re gonna get away from keyboards and, at some point, you know, typing on our phones. The sector I’m kind of bearish on is white coding. I think, you know, there’s, like, a ton of stuff, obviously, you know, platforms doing exceptionally well. It just feels like that’s a category that maybe, you know, maybe already the winners are determined. We also have to see, as these models is, you know. ChatGPT becomes a platform, and, you know, you see, you know, what Anthropic and all the other foundation models are doing, whether that sort of continues to be, you know. The hot space.
Kathleen Estreich:
Good ones.
Unknown:
Jenny Fielding I Everywhere Ventures: Yeah, I guess, building on that, Jenny Fielding I Everywhere Ventures: We talked about this a little bit, generic workflow automation for low… Jenny Fielding I Everywhere Ventures: friction tasks, right? So I guess we talked about the wrapper, but, you know, I remember when it was, like, this amazing thing to connect to off-the-shelf APIs, right, that they can talk to each other, but obviously, as that gets built into agents, that’s no longer so interesting. Jenny Fielding I Everywhere Ventures: And I think the area… so I just came back from a month in Europe, and I could tell you that where all the money is going there is into defense tech. Jenny Fielding I Everywhere Ventures: Obviously, for macro reasons, but, you know, the combination of technology and how AI’s, you know, impacting hard tech and defense tech and all the money and the government subsidies, I think, you know, 2026 is gonna be heavily about that.
Kirby Winfield:
So I think, ML Ops is dead. Long live ML Ops. So…
Kathleen Estreich:
This can…
Kirby Winfield:
or MLOps 3, this time it’s for real. Like, we’re seeing a ton of stuff around the controllability, observability, transparency, you know, evals. We’re seeing a ton of evals companies, like, I just think… again, because enterprises are finally figuring out that they can’t just throw LLMs at stuff, they’re like, oh, I have to do fine-tuning, I have to… maybe I have to build my own model, like, so I think, it should be an interesting year in that category, and I think something that’s gonna be uninvestable is just anyone who’s, like, a… like, at API, you know, brokerage layer. Like, I think that used to be really valuable, and it’s just not going to be.
Sachin Patel:
I can go. We’re a big firm, and we have investments in most categories, so I don’t want to say anything’s uninvestable, so we’ll give you two predictions. One is, I think there’s some really sleepy verticals that previously haven’t been software categories. Jenny mentioned one earlier that I think was spot on around commercial real estate. Just sort of obviously big market and a lot of ways where AI can benefit. And then second, we talked about it, I think. marketing as a category across the entire thing is sort of, like, on the cusp of something big. Image, video models are getting there. The way we discover, transact online is changing. Advertising hasn’t found its way into LLMs yet. So, like, marketing as a category, I think, is going to be pretty significant the next two years.
Kathleen Estreich:
Well, Matt, bring us home.
Matt Hersh:
Alright. Well, I’m biased, because I do mostly deals in fintech, but…
Kathleen Estreich:
I think the.
Matt Hersh:
Financial operations, and relatedly, the Treasury automation. opportunity is accelerating, especially as it relates to agents. I mean, true agent-to-agent commerce isn’t quite here yet, from my perspective. But things like, you know, fraud pattern detection that’s model-native versus rules-based is enabling, you know, lots of cool things in fintech, like… reconciliation and cash forecasting, obviously interest, optimization, compliance-related things, so, a lot of those workflows are large… have largely been human-driven over the past couple decades, right? Big teams at big banks and fintechs, and so I think there’s a huge opportunity there that we’re just scratching the surface on. And then on the commoditized side, I mean, this is probably an obvious one, but, like, the customer support in the ticketing world is… being eaten alive right now, right? I think, AI and automation and… Contextually relevant information that’s pulled from, you know, lots of different systems at a company is disintermediating the customer support. All the legacy vendors and the customer support and ticketing, and a lot of those are very large contracts, and I think we’ll see, you know, a number of those companies start to go away as more modern AI-driven solutions emerge.
Kathleen Estreich:
Awesome. Well, I think that’s our time, so good job, everyone! Thanks for joining!
Julia Nimchinski:
Thank you so much, phenomenal.
Praveen Akkiraju:
Yeah.
Kirby Winfield:
Thank you.
Julia Nimchinski:
Thank you, Kathleen. How can we support you? What’s the latest and greatest? Do you have a substack?
Kathleen Estreich:
Oh. No, I do not have a substack anymore. I had one for my previous fund, but yeah, if you are a founder building a company, come find us at Pair, pair.vc. So, investing, we’re a generalist firm, investing, kind of at the earliest stages. So, if you’re starting something or want to join something, all of our companies are hiring, so… That’s, I think, the best way to support us.
Julia Nimchinski:
Thank you so much, Matt and Kirby. Matt.
Matt Hersh:
Thank you. I mean, the same thing. We write checks up to $5 million at SEED and Series A, so, and we invest a lot more in follow-on rounds, so… I’m also actively trying to hire for a number of roles within the portfolio, so… which are all on our website, so thank you again.
Julia Nimchinski:
Very cool. Lots of folks resonated with you.