AI-Native Growth in 2026 — Fireside Chat with Kyle Poyar & Allison Snow

Julia Nimchinski:
Welcome to the show, Kyle Poir. And Allison Snow, glad to have you back. I’m gonna do the proper intro in a second, but before I do that, let’s just do one prediction for GTM and AI for 2026.
Kyle Poyar:
One prediction for GCM AI up… I would say people start looking for how to use AI for better quality instead of just more activity. I think that last year, there was a lot of activity, a lot of outbound email, a lot of automation, a lot of tasks, but conversion rate didn’t improve. And I… and I think for a lot of companies, they didn’t see that much more pipeline as a result. And so we’re now going to look at things like, how do we get even better at things like our personalized onboarding emails, because of what we can do with AI. And a lot of the higher quality work means better data and a lot more experimentation, but I think it creates a much bigger unlock.
Julia Nimchinski:
Fingers crossed. Allison, how about yourself?
Allison Snow:
I think we’re going to see the words AI and empathy and human, spend more time together in headlines. I think that this has always been something that’s… I actually think it spiked a little bit in 2016. We used to talk a little bit about this at Forrester, the idea of, understanding what folks need and being able to deliver it at scale with AI. But just thinking through. how we’re just more sensitive to customers’ needs. They’re giving us more signals than ever, we can accumulate more of them and understand them with AI. How do we respond to them, you know, not in ways in, like, batches, like we did with automation when that arrived, but in ways that truly, truly take advantage of the tech to be Far more human and empathetic.
Julia Nimchinski:
Awesome. Well… Allison Snow, welcome back to the show. Allison is the founder and managing director at Solacia Go-ToMarket Advisory, and Kyle Poirier. I think every… almost every community member is a huge fan of your, newsletter, reading it myself, every week, or even more frequently. So, growth unhinged. I heard you’re taking on Gartner. Tell us more.
Kyle Poyar:
Well, more like trying to be like Gartner, or the promise of Gartner and Forester, but for the startup world, in particular B2B startups. So trying to be that objective source of information of what’s happening in the industry, and, you know, how you can improve your performance as a company.
Julia Nimchinski:
Awesome. Allison, the stage is yours.
Allison Snow:
Awesome. I think that as far as taking on those entities, the growth of your newsletter is a really good, positive indicator of your path there. I know that Julie just mentioned Growth Unhinged, but if you’re not following, there’s a whole lot of stuff that you’re covering, Kyle, so it made it difficult to think through 30 minutes with you, but our role here is to let folks learn as much as they can in that 30 minutes, so we’ll go pretty fast. For those of you not following yet, the Growth Unhinged newsletter, even just today, I think you published something about retention for AI-based apps, just this morning, and I thought I need to plug this thing again, so… on it.
Kyle Poyar:
I… I did, with the headline being, AI companies do not have the same retention profile of SaaS, I pulled data from about 3,500 companies in ChartMogul’s dataset. ChartMogul’s a SaaS metrics and growth platform, and AI companies had a median retention rate of about 40% on an annualized basis. can compare that to more like 70% for a B2B SaaS company, and it is very different. AI companies in particular that sold for under $50 a month, so a lot of the more consumer or PLG AI companies, they are hit the hardest with this retention challenge, so their median retention for those businesses was 23%. It’s really hard to call that recurring revenue if, you know, more than 75% of your revenue leaves Within a year.
Allison Snow:
Yeah, you used the phrase, easy to buy. easy to churn, right? Easy to buy, easy to drop. I think that’s really important for a lot of folks to know as they think about I know when I was at Forester, every question that we got was somewhere in the realm of, how am I doing? Which is another word for new benchmarks, for what I’m up to. And, I think the folks really benefit from the analysis that you’re doing. So, awesome. I’m excited to be with you today. Thanks. So, our session’s description starts with a really bold claim about AI-native companies, and it says, they’re outpacing other firms and incumbents three to five times because they run on orchestration, not workflows. So, I’m not just going to read you the whole description, but I will say that as a founder and a practitioner, I read something like that, and I say, why? Tell me everything. What’s under the hood there? And I wondered if you could start By just letting folks know who joined us, because they’re saying, again, what’s under the hood, who’s in there, why would it be 3 to 5 times performance? What are some high leverage starting points? When it comes to adopting agentic AI that sort of anyone out there might benefit from.
Kyle Poyar:
Well, and I can… I can tell folks, the average company is not seeing 3 to 5 times better performance when they leverage AI. I recently did a state of B2B go-to-market report, and more than half of folks said that they’re using AI, but seeing very little impact from it. Many people are, you know, I think my mom uses ChatGPT. Saying we use, like, a ChatGPT is not enough to be AI native these days. If I look at where folks are seeing value from AI right now, it is around outbound use cases, around market intelligence, and around some content marketing use cases. if I think about where folks can go, the applications, if you really build, and kind of really focus on this, like, you can build some really custom stuff that is, like, quite disruptive to your company building. So I recently interviewed the go-to-market team at a company called Safety Culture, which is a fairly sizable vertical SaaS company out of Australia. They’re generating, 500,000 free sign-ups every… every year across 180 countries. and their sign-ups come from, like, construction, manufacturing, small businesses up to large enterprises. It becomes really difficult to do things like have human BDRs around that funnel across all the different countries, all the different languages, all the complexity around customers. Plus, it turns out that off-the-shelf data tools aren’t even that good at finding data on who these companies are and figuring out who you should reach out to. So they’re using AI to do AI-powered lead enrichment, where they actually call third-party data providers and use LLMs to scrape websites of signups to figure out, hey, what are the industries of this signup? What’s the company size? What’s going on with this company? They’ll also pull in things like, was there a safety violation in OSHA from this company? So really kind of custom data that’s really valuable to prioritize these leads. They then use that data to have an AI inbound BDR that has a personalized outreach, can even respond to questions it gets back from prospects, and then sets up the meeting booking for the AE. this has been a game changer in terms of productivity, but it started because it had really good data. Like, I don’t think the auto BDR would have been that productive if it didn’t have really high-quality, clean, enriched data in the first place. And then as people sign up. There’s an AI lifecycle personalization where all of the different industries and profiles of companies are matched against the top features that are going to be most likely to be used. for those folks, and why that industry needs that feature. And so they essentially have personalized onboarding journeys across every type of customer that AI enables that drives a massive uplift. in activation and conversion. And then the final thing is they built an app layer above their CRM that essentially provides the sales team with all the customer information in one place, all the past interactions across digital touchpoints and their own past touchpoints, and recommends the exact next best action, and allows them to even orchestrate some of those actions automatically. So I think the other thing that they’re doing is they’re not just having it be, like, it’s the AI for its people, they’re actually.
Allison Snow:
2 million.
Kyle Poyar:
Building tools to make people a lot more productive as well.
Allison Snow:
I love that. I also love that it sounds like that company kind of started by saying. what is uniquely challenging for us, right? And part of that is the market that is so… Yeah. Geographically diverse, and… language diverse and all of those things. Do you find that folks are typically pretty… well, first of all, do you agree that that’s a good place to start, with use cases and unique problems at a company, and then think about what tech what AI functions, what AI tools might help, and if so, do you think people are doing that pretty consistently, or do you still see a lot of folks kind of saying, that’s shiny, how do I get some in my company?
Kyle Poyar:
So, I think Brendan Short actually had a really great piece recently on this idea of, sort of, signal-based selling and using intent data. And there was a time 10 years ago when if you found out a company had just raised, like, a Series B or a Series C or Series D, like, that was really valuable information. You would want to prospect into them at that point in time, because they would be ready to buy your product. Now, there’s probably a couple thousand companies that all have that data, all get at the same time, and could set up an AI workflow to send a message to that company. You’re sort of lost in a sea of outbound. And it just makes it harder to stand out. And so I think the big thing to me is you’ve got to find ways to take advantage of AI and better data tooling that are custom to your specific businesses and the problems you’re seeing, as opposed to copying some other playbook. Because if a playbook is copyable, that means it’s going to be outdated really quickly. And so, needing to find something that’s custom to your business, and I think the other thing is you need to focus on the biggest unlocks in your funnel, and so that could be… maybe you have a really great way of closing business, but you don’t have a lot… a great way of getting intent, or high intent customers. Right? Focus the applications around, you know, that… the biggest bottleneck that you’re facing. That would be kind of my… high-level advice.
Allison Snow:
Yeah, no, sounds great. So you give very specific guidance. I want to dive into… to some of that. One is, AEO. Agentic Engine Optimization, and you have said that that could be as foundational… I believe you said, I believe it was in your newsletter, so I don’t want to misquote you, that it is as foundational as SEO was. In 2004.
Kyle Poyar:
So it will… it is getting there. So I’d say if I look at, when I’ve talked to companies. talked to a bunch of folks who told me that Answer Engine Optimization, or AEO, is essentially their number one fastest growing channel right now, and that when they look at it from traffic alone, some companies do get a good amount of referral traffic. Webflow, I think. You know, they told me they got a pretty healthy, actually, amount of referral traffic specifically from these engines.
Allison Snow:
and even.
Kyle Poyar:
or specifically from ChatGPT. But for most companies, it actually doesn’t show up as, like, there’s a referral link, and someone goes from ChatGPT to your website. what happens is they are asking questions to AI search, getting answers, and essentially having their buying process informed by AI, and then they show up much more ready to buy. So the traffic that comes from ChatGPT, or just AI tools in general, is often, like, much higher converting than any other traffic, because they’re doing a lot more of the buying journey on AI tools and elsewhere, as opposed to coming to your site. And so you… and you also, you know, you can’t really tell that this is ChatGPT, because they will usually maybe come to your site directly, because they now know your brand, or they ask Google about your brand, and then it’s a branded search. And so you might track it in your attribution systems as, like, Google or a direct source. But if you ask the customer, hey, how did you hear about us? If you ask them consistently, you’ll probably find that, you know, 10% or more of that is actually already ChatGPT, and it’s growing 3 to 4X year on year, and so if these trends continue going into next year, many companies might see 15-20% of their, you know, high-intent, demo requests coming from these kinds of channels. And also, this just makes it a… really important source of demand right now, I would say particularly for categories where there’s already a demand that exists in the market. And so, if you’re in, like, a website building space, that is an existing category with a lot of existing search behavior, and the search behavior is just switched from Google to AI tools. If you’re in a category that, like, doesn’t really exist, like, you’re the first person doing the thing that you’re doing, there’s not a lot of search intent around that category, and so it’s not going to be as effective for you until you can educate the market to try to buy these things. But when the market does change, it does get educated, you want to be the brand that shows up first.
Allison Snow:
Yeah. That’s really actionable, thanks. Any tips on being that brand that shows up first? I think a lot of people in the audience are sort of worried about being invisible on that channel, and 10% to 20%, that gets pretty scary real fast.
Kyle Poyar:
So a lot of it, you know, I’d say initially comes down to having a really defined set of prompts or topics that you want to be known for. And so, maybe if we go back to, well, another company I featured, their AEO journey was a company called Techevo, which is an enterprise-grade learning management system. Now, you know, they love to be the number one tool ChatGBT recommends anytime someone’s looking for a learning management system, but, like, the reality is. they have some use cases where they’re much stronger, they have some customer types they’re trying to go after. They’re generally trying to go after enterprise customers, and they have certain differentiators, like multilingual, where they want… anytime someone has that need, they want to be the number one vendor for it. So I think the first point is. having really specific topics that you think you would want to be known for. And so if you’re a shoe brand, like, you don’t just want to be known for shoes, it’s, like, running shoes, or is it cushion running shoes, or durable running shoes, or low-price running shoes, like, you need those modifiers as well. So it starts with the strategy piece, and that’s something that’s hard to outsource, hard to use AI on, because if you pick the wrong things to be known for, then you’re going to attract the wrong traffic, right? But, so start by picking things you want to be known for, and then it comes down to building content and getting your brand known for those things. And the first things you can do is just build up content on your site. These AI tools love, like, listicles, they love comparative content. It’s a little awkward to write about yourself versus your competitors. But AI tools actually really want that data, and they’ll ingest that, and then often kind of plagiarize it back to other folks.
Allison Snow:
But that means you need to have these pages be built out, be very detailed, ideally include mostly text and tables as opposed to images, because AI prefers text and tables over images.
Kyle Poyar:
You want it to be structured in a way that’s easily digestible for AI, so a lot of different sections, clear headers, a frequently asked questions section, right? This sort of Q&A style is really easy for AI to digest. And you want to keep that fresh, because these AI engines want to find the latest and greatest content. And the final thing is, you want to create surround sound around that content, so ideally, other places are citing your material, your brand is being mentioned in places like Reddit, on LinkedIn, on user-generated content sites, because you need that surround sound to add additional credibility. But one kind of hack that I’ve seen some companies take is actually partnering with or even buying third-party media brands or community sites, and partly because these brands are already seen as objective, sort of trust-based sources, and so if a software company buys a media asset, this actually becomes a really great place for a lot of their AEO content to get, sort of, trusted from LLMs, and then bring them back to their software company.
Allison Snow:
That’s awesome. People… people love hacks. That’s a smart one.
Kyle Poyar:
It can be an expensive hack to buy a media company, but it could be worth it if you can monetize that traffic well enough.
Allison Snow:
Yeah, it’s, you know, you’re saving time what you spend in money. There, I imagine. Cool. Staying deep here, auto BDRs. AI-powered BDRs, they seem to be a pretty early agentic success story. You, you mentioned, them earlier. Are you seeing good… good examples of this?

Kyle Poyar:
So I’ve been thinking about this… A couple different ways. So there’s… there’s outbound, BDRs, SDRs, and then there’s inbound. On the outbound side, and then even, you know, within the outbound side, that you could buy an off-the-shelf AI SDR and kind of just fully outsource it. you could try to automate certain plays, or you could try to enable your existing, kind of, human team. And I would say. there was a lot of excitement around that fully outsourced model, the AI BDR. I’m hearing a lot of, sort of, disillusionment around that right now. And part of it is, you know, people didn’t really have that much of a strategy turned to outside vendors, hoping that those vendors would have some sort of secret sauce that they could benefit from, but they didn’t see that, right? I think there’s a lot more success right now in the more automated plays approach, and specifically looking for highly specific kind of signals that a customer is a good fit for your product, or is showing some side of… some sort of buying intent. So maybe they’re visiting your website. But they didn’t… request a demo. Well, you could have a kind of an automated message that feels a bit personalized that gets triggered when that happens. So if you base this sort of… these automated campaigns on first-party data that others don’t have that you have, or where there’s some sort of signal that you’re getting around a connection to your product or your brand, or your category. the better the signal, the easier the message is. The message kind of writes itself, if you have a clear enough signal, and so you don’t need to do as much educating. It’s more about converting the demand that already exists out there. You just haven’t captured it yet. So I’d say, like, that’s a really powerful… use case, and then that applies to the inbound BDR-SDR use case as well. Where for a lot of the inbound, it’s about building, you know, a mechanism to qualify accounts, and qualify their interests before passing them to an AE. And so, for these folks, you know, usually if you were doing some sort of marketing scoring, lead scoring, you’d be using pretty generic data. I think LLMs are actually much better at lead scoring than people expect, or would, like, give them credit for. And then the kind of inbound BDRs and SDRs can kind of confirm that interest, and schedule the meeting right past folks to an AE. And so I’d say that’s more of an efficiency play, but can be a powerful efficiency mechanism.
Allison Snow:
That’s excellent. I think it’s a good reminder, too, right off the top, when you mentioned that sometimes the outsourcing, it’s not so much that the outsourcing didn’t necessarily work, but a lot of folks maybe said. I’m going to outsource the full piece here, strategy plus the outbound and the execution, and that’s not really something… That reliably works. It’s a good reminder.
Kyle Poyar:
Yeah, I’m finding, you know, I go back and forth on this, because for a lot of companies. I’d say tech companies love to build their own stuff, and many, I’d say, overbuild. Because, like, I’ve met and worked with many companies that even, like, have built their own CRMs, because they think there’s some sort of, like, special thing that they need that they can’t exist. I’ve seen people kind of custom build their own versions of Calendly and other tools. Custom build and type form. Often those resources are better spent building things that, you know, you can ship to customers, as opposed to just internal stuff. But there’s just always a bias within tech to build. I think in AI, though, that bias is right, and you can actually have a leg up over competitors by building your own custom solutions That fit your workflows, and often you want to maybe do something manual to figure out what is that process or workflow that works best for you, and then use AI to automate that, or increase the scale at which you do that. And you’re not going to really be able to run that playbook if you just try to copy and paste someone else’s playbook, or just bring in a third-party tool without knowing that process that you want to run in the first place. So I do think that now is actually an interesting time to build internally.
Allison Snow:
That’s really interesting. I think people will really benefit from that piece of advice. some… really puts sort of common wisdom on its head. I think tech companies are kind of overbuilders, and, bias might be, I’ll stop doing that, but it’s really interesting to kind of look twice and say, there are exceptions. Very cool. Speaking of exceptions, when we think about pricing. It looks like there are some products, like, sort of AI products that underperform because they’re priced like SaaS. Do you agree? You don’t have to invest. Give people what they want, a little controversy, a little argument.
Kyle Poyar:
Well, I’d say, like, we’re all trying to figure out pricing for AI products. I don’t think anyone has cracked the code on the best fit model. If you copy a lot of what’s been done in SaaS, for many AI companies, that is going to be particularly, like, putting you in a particularly tough place. And so, for example, if you have a seat-based model, and your customers can use the product in an unlimited manner, and you have full access to the latest and greatest models, you might find that your best customers are extremely unprofitable. And so for many companies that have sort of flat fee subscriptions or seat-based models. It’s predictable for the customers. Customers often like buying that way, but it leads to, like, a significant risk of unprofitability, or customers that are unprofitable. There’s not as much upside potential, especially with, kind of, flat fee subscription models. And you’re at risk if your customers are really successful with your product and need less headcount, then they’re not going to keep buying more licenses, right? And so, as your product works really well for them, you actually make less money, which I don’t think the position that anyone wants to be in. That said, like, there’s certain cases where, like, I still find seat-based models work quite well, and there’s a lot of companies that are trying to get creative with seat-based models, where there’s kind of a hybrid model. Where you actually have a seat-based monetization mechanism, and then an amount of usage that’s included, and then an ability to expand based on usage above and beyond that. Where there’s kind of a variety of mechanisms that I lump all within hybrid pricing that essentially give companies more flexibility around how they land and expand customers. And then, you know, I think on the bleeding edge, there are some places that have gotten really far away from typical SaaS pricing, and they’re experimenting with things like outcome-based pricing, particularly in the AI customer support space. I think that’s a very exciting set of activity that’s happening. I’m following that closely, but I’m not seeing as many companies being able to pull it off. as I thought, like, Intercom was an early adopter a couple years ago. A lot of the players in customer support have followed suit, but it has been very slow to expand beyond customer support.

Allison Snow:
Very interesting. Is there anything that… I feel like people are kind of… maybe I’m wrong, just thinking, oh, I’m not AI native, I’m SaaS, and there’s these kind of walls, between what those folks can do, and I’m not even sure what AI native means, I’m sure it means a little something different to everybody. But are folks in SaaS, in your observation. sort of copying anything from AI-based pricing and being successful at it? Or is there anything that they can learn if they’re kind of watching and saying, there’s nothing about my company that’s AI-native, but gosh, if there’s a pricing tip for me, I’ll take it.
Kyle Poyar:
So most… I mean, most companies at this point are hybrids in some way. I don’t meet that many SaaS companies that don’t have some AI capability that they’re offering to customers or selling to customers. And so, I do think that, And, you know, actually, SaaS companies are trying to become more like AI companies, and AI companies are often saying, hey, we had AI, but now we need to build more integrations and workflows, and, like, we need to be stickier within our existing customers, and SaaS is really sticky. And so, AI companies are actually trying to become more like SaaS. But I think the big thing for estimates to take away from AI is that, there’s a really big potential to, deliver a lot more value to customers, and what I mean by that is AI products, especially when combined with services. can kind of take on whole work products. So instead of just being a tool that people use to be more productive, it can do the work itself. And that is a value proposition that’s, like, potentially 10x stronger than just being a tool for you to do the work. headcount budgets are way higher than tech budgets. You can also tap into agency or consulting spend, which are often massive amounts of spend. And so, if you can understand how your product can actually do the work on behalf of your customers, and how you can make sure that you have a consistent way of measuring that and getting attribution for it. I do think that, like, that is what unlocks a lot of pricing power, but you need to be able to solve that bigger problem and have attribution for it in order to unlock that.
Allison Snow:
Excellent. Excellent. I know that Julia is probably going to start her Process of seeing us out, as we head into the half an hour.
Julia Nimchinski:
Incredible session. Thank you so much, Kyle. Thank you, Allison. Took so many notes. Last question. What’s the best way to support you? Kyle, is it the newsletter? What’s next? For you, for.
Kyle Poyar:
Yeah, it’s subscribing to the newsletter. It’s Growth Unhinged, you can find it at growthunhinged.com. I publish once or twice a week, and a ton of good content with kind of practical advice around how to implement AI, how to use AI for go-to-market, and then just stories around how the fastest growing companies are succeeding.
Julia Nimchinski:
Amazing. And Allison, how about yourself?
Allison Snow:
Very cool. I will, just quickly double down on Kyle’s newsletter. He does a ton of, real-life examples, which I think is helpful. I also have a substack. There’s 3 so far. I plan to publish once a week. Find me, Allison Snow, Celacia GTM, up there, and, support me if you’d like, and as always, excited to be here. Thank you.
Julia Nimchinski:
Huge pleasure. Thank you so much again, and we are transitioning to our next session. Welcome to the show, Kimberly Storen, Chief Marketing and Communications Officer at Zoom, and Angela Winingar, Head of Marketing at Invisible Technologies, and we have Eric Charles.