What the next gen AI GTM tool should look like with Unify's Austin Hughes

Yeah, today, in a variety of forms and factors, we actually drive close to half our revenue from our own product. Wow, that's incredible. So it's really, you know, it's obviously a great testimonial and it shows in a bunch of different ways, but it's also, I think, really important for us because we're also the biggest critics of ourselves in doing that.

Hi everybody, this is Tetsuro. We're here with another episode of our Off Topic video interview series. I am very excited to be here in San Francisco at the offices of Unify. Alongside Austin Hughes. Austin, thank you for having us.

Thanks for having me on the show.

And for the viewers who don't know you, can you give a quick intro about who you are and then what Unify does?

Definitely. I'm Austin, co-founder and CEO here at Unify. We're building the future of go-to-market in a world where software does a lot of the work for sellers. And I can share a lot more about what that looks like in a second, but we're Venture-backed company, Series B stage, have raised about $60 million to date and are backed by folks like OpenAI, Thrive, Emergence, and Battery as well.

And so how did, you know, let's dive into the product first. How did you come up with it? And like what does it specifically do when you mean, you know, go-to-market?

Yeah, definitely. So the genesis of the company was back in 2023, we got started. And at the time, the software markets were terrible. It was very difficult to sell, valuations were terrible. And the number one concern in board meetings we kept hearing was that pipeline was impossible to drive. And so for folks who don't know, every new customer conversation in B2B starts with a touchpoint that says, hey, we should talk and I should explain to you what my product does. And so we view that as pipeline. Our job is to help create that pipeline for companies and to serve as a thoughtful partner, in some cases, kind of like a forward-deployed engineering team of just really strong technical folks who think about go-to-market problems. And so we spend a lot of time thinking about outbound and outbound sales in particular, but that's our domain.

And it's— I've worked in VC as well, and so talking to a lot of SaaS companies, oftentimes they've built a pipeline, but sometimes it's not real pipeline, it's like they've built a touchpoint of a lot of people, and then they don't know how to turn it into action and insights, right? And how does the unified product lead people from not just from pipeline, but to actual customers?

It's a good question. We start actually, the difference in our product and how it's quite different from what else is out on the market is that we believe in right person, right message, right time. And so, uh, the analogy that I like to give is if you're walking down the streets in San Francisco, it's very hard to sell someone a house in Arizona. Uh, it's just not somebody who's oftentimes looking, and why would they have a reason to want that product? Um, and our job is to help bridge the gap between our customers and their customers, and it's actually a search problem as we see it. So we So we are searching for the prospects who actually are trying to solve the problem that our customers solve. And so, when we view it from that lens, it gets back to this question of closing and actually being able to turn those leads into customers, because we're focused on the people that actually have the problem that end up turning into healthy customers that evolve from there. So it all starts with the targeting.

[Speaker:AUSTIN_HUGHES] So how do you actually kind of find those people, and how do you know that they have the problem? Because oftentimes, maybe they're not even in the database of these customers. Activities of your customers, right? Mm-hmm, mm-hmm.

So I would break down the world, it's all about data and how you work with data. And in our world there's two buckets, it's third party and first party. So first party is data that is close to your actual product service area. So maybe it's your marketing website, maybe it's actually you have a free product and someone's logging in and doing actions there. So that would be first party data. And then third party data is things that you can observe outside in that are not necessarily things that your company has unique access to, but they correspond to those value moments. And so yeah, as an example for us, something that might correspond to that is, Hey, somebody just hired a new marketing leader, as an example, or they're bringing on a bunch of new sellers, or maybe something else, a new exec started in a specific title. That tells us that this company is investing dollars and time and resources towards the roles that are driving pipeline, which means that what we do is top of mind. And so, combination of those two first and third-party buckets means that we're able to search through a huge portion of the TAM to actually find who to reach out to.

So I think you started Unify originally. You worked at Ramp previously, and I think you had found some issues around cold email efficiencies. I think that's what the article mentioned. Like, what does that mean, and what kind of led you to take this into a completely separate product?

Yeah, totally. So Ramp was an incredible experience for me. I joined as an early employee. I was there at about employee 40. And it was an interesting challenge because it was a very competitive market. You had a market that was huge, but it was owned by many incumbents that were extremely well-funded, many public banks, things like that. And the biggest problems we faced were distribution problems. It was how do you get to market, and how do you actually get people to care, right? So I think in that world where you've got millions of companies that you can sell to, you have a very small team, you just have to think outside the box if you really want to break through in today's day and age. And so I started spending a lot of time thinking about, the world of outbound from more of like an engineering-led perspective early on, and always thought that was fascinating. Wound up to where we are at Unify because very consistently as we started the company, as we would talk to executive teams, founders, didn't matter who it was, we wanted to build something that great startup founders loved. Pipeline was just consistently the thing that kept coming up as an issue and had particular domain experience in outbound and felt like a natural fit also for all the technology that we were seeing being released. ChatGPT came out, call it like 8 weeks before we started the company, which was serendipitous timing.

And I think we should talk a little bit more about this problem about distribution because I think it's something that a lot of people don't realize, and especially in Japan as well, where I think a lot of people, there's so much focus on product and functionality and products that a lot of people, even in Silicon Valley, I think if a product works, it'll just sell. And that's true to some extent. But I think Peter Thiel and a bunch of folks have really pressed the big need of having great distribution. Can you talk a little bit about the issue that we face, and especially in today's age with AI?

Yeah, definitely. Distribution is just a hard problem space. And you wind back the clock 20, 30 years ago, it was very difficult to build software, right? If you wanted to go stand up a new piece of software, you had to go actually go to a data center. You had to stand up your own servers. And it was just a very different world. Today, that plus so much more is on demand. Right? Even something today as simple as, hey, we want to stand up a new system for billing or for notifications in our product. There's infrastructure out there that does that for you. And so it's so much easier to build a company, and it's so much cheaper too as well. With a few thousand bucks, you could build a production-grade piece of software if you've got the right engineering talent behind it. And so that's just the world we live in. And because of that, you also have the capital markets are extremely well developed now in the US in particular, right? So there is so much competition being created all the time. This is a ton of new creation of companies, and And by and large, I think that's a very good thing for the world, and I'm excited by that. But I also think that it means that it changes the game on the field, and you have to be aware of that as you're selling software. And so, you know, in our category, there's hundreds or thousands of companies that we'll compete with in some form of capacity. I think what's interesting is that sort of changes as you get to different levels of credibility and who you actually see as competition. But the most important thing and the biggest reason why companies never truly break out is they can't figure out how to get enough attention on their product. And it's a challenge, right? Because in the early days, you don't have something that's that exciting, and you don't have anything that people are really raving about. Maybe you got one feature that people kinda like, and it's about how do you manufacture the excitement, how do you build momentum? And really, that comes down to, that's a distribution problem at its hand. And for us, early days, actually, funny enough, being really active on social media was the way that we actually started to build credibility and excitement about what we were doing. And before that, it was incredibly hard to, I would go up to somebody at a conference, be like, hey, I'm at Unify, and they're like, what's Unify? Today that actually has changed quite a bit, and most people I talk to have heard about us in some capacity, so it's really nice.

And I think especially most recently, right, even in San Francisco, you see a lot of camera crews around, because everybody's doing launch videos now, right? And previously you would never do that. You would just try to go and get on TechCrunch, or you would go on like Product Hunt or Hacker News, but now you've seen you need to step up the game in order to actually get some attention. So I think that's what you guys are doing, kind of the next step beyond that, right? It's not just the launch side, it's afterwards.

Yeah, exactly, exactly.

Makes sense. So you list, who are some of your customers and how do they specifically use Unify?

Yeah, we work with a ton of incredible companies. We largely serve companies today that are B2B software companies. It's sort of evolving in real time. We've got 4 or 500 customers today, and so it's evolving quickly, but today we'll work with folks like Cursor, Perplexity, Flock Safety, a bunch of other incredible names, but a lot of these sort of marquee software companies that are coming about.

And do they use it specifically, like, is there a specific way they use it? Is there, yeah.

It's a good question. So it really depends on the company, and it all comes down to this idea though of finding the right people to be reaching out to in an interesting way. And so maybe a couple examples is, you know, Somebody like a Cursor or Perplexity or a company that has more of this sort of like PLG and like high demand business model, we'll try to tap into that. And that's sort of the strategy is, hey, help companies to take that data. The main problem that they face is how do you act on that and turn that into automated outreach at scale? So we'll work on that problem with them. Somebody like Flock Safety though, it can be quite different. Like as an example, one, one thing that we do with them, Flock Safety is a security company that their mission is to make communities safer. One of the things that has worked really well for them is actually the most compelling event for why you buy their product is some sort of crime happens, right? There's a break-in near a location, or you have a store that gets vandalized, whatever it might be. And we actually send agents out every day to help them to detect when new crime events happen near one of their customers' locations, and we use that as a reason to reach out, and we actually use that as a reason to what to say, and that works incredibly well. In that case, we're using agents and it's very different from the first-party data I was just talking about.

So for example, in that scenario, do you work with Flock Safety in terms of figuring out what specific moment in time that customers tend to convert and then you specifically cater agents to that problem, or is that entire process almost automated as well?

Really good question. Today, that's done a lot by our team in partnership with their team, is thinking through that strategy and saying, Hey, as a creative marketing team, what do we think are the reasons for each route? But I will say that we're working on turning that into much more of an AI-driven problem, and the results have been pretty interesting. And I think we're not far off from that.

Yeah, I mean, I guess it's understanding when and how your customers find you. And so as long as you can figure out a way to automate figuring that part out, whether it's asking the clients themselves or going directly to customers, then you can translate that into events.

Exactly. AI also just, it can know so much more about so many sub-industries than any one human can. You have agents there, AI can beat the best people at math in the world, different problems. The same is actually true in the world of go-to-market. We have these experiences now where we help companies to identify signals as they move into the product for the first time. The signals and the insights are so much more granular than anything that our human team could come up with because the models just know different industries extremely well. And so I think that we're going to move towards a world where the agents are actually much more responsible for the creative thinking on top of just the execution. And we're starting to see that happen in real time.

What would be like a good example of that? And yeah, of that kind of granular kind of detail that an AI would realize that maybe a human would kind of miss out on?

We have a couple companies that sell into the biotech space, as an example, and something that we'll see happen is the way that you sell into their customers really depends on like things as granular as the type of manufacturing facility that their customers have. And the manufacturing facility, it's not like, hey, like it's something specific as like the certain subtype of like bio. It's stuff that's well beyond my realm of expertise, but There'll be things like, hey, these 4 subtypes really matter in this industry, and this company has this type of manufacturing facility, and they have 2 of them, as an example. And that's something that I or a post-sales team would have to do quite a bit of research to come up with, and the model just knows it out the gate. And it's funny, the customer reaction the first time they come into a product like that, they're like, you have the best forward-deployed team there ever is, and we're just like, it's not, it's AI actually. So it's cool to see that happen.

Yeah, that's really interesting. And so there's a lot of like AI sales tools or like GTM tools out there in terms of kind of differences between what you guys do and them. Like what do you think are some of the differences?

I think the main difference is that we early on were big believers that to do interesting things in the world of sales, you had to solve a bunch of really boring and painful problems first. And for us, that really stems back to CRM. So Salesforce and HubSpot, data hygiene issues, it stems back to dealing with duplicates, it stems back to things like rules of engagement and understanding who to go after and who not to. And in our world, you have to build all these very sort of important but not that interesting primitives to be able to do really interesting things on top of. And so, I think that our differentiation today really stems from the fact that we believed that early on and we took that challenge head on. As an example, one of our founding engineers, who's incredible, spent the first 3 months of his time at Unify just building our Salesforce integration.

Oh, wow.

This is before we had any customers, and that was sort of the first major thing that we built. And it's just one example of, you know, what that looks like in practice. But we had to do a lot of things that were sort of like that along the way. And I just don't think that many companies in our space, they've sort of gravitated towards the AI problems first because they seem most interesting to work on and more fun. And that is true to some degree, but the problem is that AI is only as useful as the data that it has access to. And to have good data in go-to-market, You actually have to solve that from an integration-level perspective first. And so that's sort of what we've bet big on, and so far it's been paying off.

So speaking of companies like Salesforce, HubSpot, all these kind of CRM providers, they often kind of describe themselves as system of records, right? I think you've been pretty clear in saying that you're not just— you're not a system of record, you're a system of action, right? What does that actually mean, and how is that different?

So, you wind back the clock 20 or 30 years, and the CRM, as, you know, back in the early 2000s, it was simply a place you could come and write someone's name or someone's email or contact information in a system, and you did that manually as a human. We had a big shift maybe 15 years ago where we started to move everything to be much more programmatic, where instead of manually keying in all these records, you might integrate in different systems that would help bring in some of those records, you'd start to manage some of this data, data and scale. Salesforce did an incredible thing around that time, which was actually to open up their platform in a way that I, you know, I wasn't thinking about CRM software at the time, but I have to imagine that was fairly counterintuitive, is how do you, you know, build lock-in? You open up your ecosystem so that anybody can build on it, but it created a tremendous amount of lock-in over the past, call it, 10, 15 years, where people will build these incredibly complex systems on top of CRMs that feed data in, and they're business-critical, right? There's information flowing frequently, and that information is worth a lot of money. And so that's sort of the paradigm that we've lived in. And I think AI really switched or flipped the switch on, on what software is and what it's expected to do. And our perspective is that over the course of the next 10 years, what's really going to happen is a lot of these systems of record that got their sort of moat and lock-in from being integrated and being stores of information, that's not going to be good enough in the next decade. And what's going to actually happen is The default mode of a piece of important software is gonna be that it can do an incredible amount of work on behalf of you, and it can be incredibly helpful at helping your team get more leverage. So the way we think about it is, how— what would the world look like if actually 95% of the really boring and monotonous tasks that sellers have to do, they could do through our software? As an example, the last 5% that are really interesting, they can actually spend their time working on. And our perspective is that this system of action, as we call it, this is something that just does all this work for you, for you is actually gonna become the next iteration of whatever we call the CRM category today as the core system of work for go-to-market teams. And so that's how we see it playing out. I think there's a lot of different takes on how you get there and what the steps look like. We have our own perspective on that, but I think what we're gonna see inevitably over the next 10 years is that we're going to see a new system emerge as the core system of work for go-to-market teams.

[Speaker] So I'm assuming that entails some sort of rethinking of the entire system itself, itself, right? Where to even start from. Like, where is the difference in terms of where it starts from?

So you have to provide, in building a company, you have to provide some sort of differentiation up front, I think, that helps get people interested. For us, a lot of that was in this idea of being able to turn signals, whether they're found by AI, whether it's data, whether it's first party, third party, and turn that into action and actions against customer that actually create creates revenue. So, um, that was sort of the interesting, uh, sort of wedge for us. I do think though that what becomes really important and the thing to look for over time is that there's people coming at this problem from different directions. A core system of record has a ton of data reading and writing out of it. I think that's what you want to sort of look for is how can somebody be thoughtful and be creative about getting those writes and reads in and out of the platform and make sure that you are becoming that system. And so my perspective is that coming directly at the CRM category is probably not the interesting thing to do over the course of the next decade, because you're going to be going where the puck is, not where it's getting to. And our perspective is that we're building something that's somewhat adjacent, that serves a different need, but over time, the problems foundationally, the data primitives, et cetera, are all very similar, and it's something we can broaden into.

Interesting. Unify uses its own product to partially dogfood it, but also to prove out that it works as well. Can you talk a little about, is there any specific way your team uses it and how much of it is being used right now?

Yeah, today in a variety of forms and factors, we actually drive close to half our revenue from our own product. Wow, that's incredible. It's obviously a great testimonial and it shows them in a bunch of different ways, but It's also, I think, really important for us because we're also the biggest critics of ourselves in doing that. And I don't know the exact numbers anymore, but at one point we would submit more bugs and tickets internally than any of our customers would individually, and even in aggregate sometimes. And so we're big critiquers of our own work. And it also just, I think dogfooding your own product just makes for a better customer experience because there's certain things that you feel as you're using things that just don't feel quite right, and they're emotional and they're a little bit hard to explain, and you see when you're actually using it. But the ways we use the product today are both in a fully automated capacity, so being able to do things like take website traffic of ours and retarget it with automated outreach, all the way down to our reps, actually it's their core system of work for doing outreach. And so whether it's them sourcing one-off folks to go after, or it's, hey, we detected a ton of intent on this high-value person, let's tee up a phone call task, all that runs through Unify today. And so, yeah.

Got it. And I'm assuming on the product, you can kind of choose what you want to automate, what you want to kind of have a human actually look at, what do you want actually kind of very, very manual as well. So each company can kind of set that as they want.

Yep, exactly, exactly. We started with the automated approach, then we built the human component.

Got it, got it. And then do you make recommendation in terms of like how much to like automate? And does that kind of slowly grow? Over time, or does it actually grow really fast and then slowly lessen to some sort of standard level?

We do. In go-to-market, it's interesting though. It's a little bit less of a rollout process based off of comfort level, and it's a little bit more of what's happening at that company in that business. So as an example, if someone sells into the Fortune 50 or enterprises, usually there's different targeting strategies, and you want to pull a human in the loop a lot more on that outreach because each touchpoint really matters. The phone is also really gonna matter a lot, and humans are better than AI are today, and so sort of depends, like, as in selling into the enterprise, that's how you think about it, but if you're selling into SMBs and more commercial businesses, oftentimes there is a crazy amount of your TAM, your total addressable market, that's just untouched, and a really easy way that we slot into a team is to say, hey, of those, you know, 50,000 companies, you actually should be targeting about 5,000 of those a month. You don't have the sales team to do that, though. You should stand up Unify as a sales team system to help you engage with them. And that works really well as well. So it sort of depends upon the segment someone sells into.

Got it. There's a critique in the industry where, because there's a lot more AI sales products, that every cold outreach email looks the same, or everybody assumes it's AI. Have you actually found that to be the case? And if that is the case, how do you solve that?

I think it just goes back to the— one, I do think it's a problem. And I think it's a problem that— you know, has gotten— it was bad before AI. People would just send a lot of spam emails. I do think that at its core, if you're focused on reaching out to people who have a core need for your problem, they don't care a whole lot about whether it was automated or whether it was written by a human. I'll give you an example is I bought a product from a vendor, Carta actually, and I typically keep a to-do list of my top 2 or 3 things to do every week. At one point back in early days of the company, one of my to-do to-do list items was to find a way to communicate to our employees the value of stock options and how to think about them. There's a lot of complexities around tax and different components like that. And the to-do I had for myself was to write a doc that kind of explained this, but Carta actually reached out to me sort of spontaneously and said, hey, we have this product, it's called Carta Tax Advisory, it's geared to help employees understand options, et cetera. They get access to a personal tax advisor, and would you be interested? And I think I bought the product like 4 hours later. And so I have no idea if that email was actually automated or personalized, but it didn't really matter to me. I had a problem that I was trying to solve, and it was at the top of my list, it really mattered, and that solution sort of fit seamlessly into that work. And so I think ultimately at the end of the day, if you're reaching out to people who have real problems, the rest starts to fall into place. To fall away. But I think the problem with outreach today is that 98% of it is not to people who actually have problems. And so, yeah. Yes.

So we need way more targeted kind of like—

Yes, exactly.

Exactly. Makes sense. Makes sense. And then I remember reading that you intentionally kind of made the decision to not fine-tune your kind of AI models. I think a lot of founders would say differently that that's part of their differentiation factor, right? That they have, you know, very few people can copy them because of that. What was kind of your thought behind this?

Our general belief system is that the foundation model companies are really building state-of-the-art and are constantly sort of shipping and iterating on that. And, you know, not only that, but also the primitives around AI to make it really useful. And so as a product team, our job is to deliver the best value we possibly can for our customers. And it doesn't really matter if we do that work or if someone else does it, et cetera. But we have to provide the best mechanisms. And one of the things that I think we've been fairly opinionated on is that the foundation models are going to keep getting better and we want to bet on them continuously getting much, much better. And one of the best ways to do that is just not invest that heavily beyond how you fine-tune the models and things like that and assume that they're going to get better at most tasks and there's going to be simple frameworks, et cetera. So I would also say that we do do some— you do work with a variety of model companies and do some work to make sure that the model is working well for what we're doing. But I will also say that we try to be very concerted when we do invest in customizing anything to make sure that we are doing so in a way that we feel like it's defensible and it's going to continue to stack advantages over time. So we try to be pretty methodical about that.

And you don't have to immediately change when OpenAI releases their newest model or Anthropic releases their newest model.

Exactly.

Makes sense.

We're big believers in moving as quickly as we possibly can, and we really try to avoid throwaway work.

Got it. You're finding differentiation in other places, whether it's the workflow surrounding it or—

Yeah. Yeah, exactly.

Got it. Then back in April of this year, your team built a tool for companies to see how they're ranking on AI search, which I think is a great tool and I think you provided for free.

Yeah.

Explain what this was and then whether, one, it had an effect and two, do you think more companies will use software almost as like content, right?

I hope so. So, you know, I think ultimately you can think about marketing. We try to be— we are at our core a company that really thinks about distribution and unfair advantages, and so we try to push ourselves to, you know, to be creative in that sort of way. And so I think that you can look at a lot of things in the context of marketing, right? And I think this is an example of that. We built what would be considered TikTok, typically a product, but we gave it away for free, and the reason why is that we actually felt like it was a great marketing tool for our ICP, for people we wanted to talk to. And so, we do that with a variety of things, but I think more companies should do that. I think it's, you know, there's the traditional channels of like buying ads and things like that, like I think, you know, they can be done well, and they can be done interestingly, but I think also there's a lot of room for creativity to push the eyes of like, how do you get attention in this world? I think you'll notice this even with PLG companies as an example, where they'll consider the hosting costs and things associated with their free users actually as marketing expense, which is sort of an unorthodox way to think about that. And so I very much view this in the same way as what are all these interesting bets of things we can do.

Yeah, no, it makes a lot of sense. Talk about what you call the observation model. What is it? How does it work? What is it, first of all?

Yeah, so the observation model is a piece of our core product offering. And in particular, it is a way to think about how AI in our product should work. And it's a somewhat big break from the traditional way of the world. Most products that you build today, and this is still true, they're built off of relational databases. And there's data associated with different entities and things like that. And that's a very good way to store a lot of information. But I do think that in the world of AI, you're going to start to need different paradigms to think about these things. And one of the ways that we've done this is with our observation model. And the idea is to try to think and act in the same way that a human would and human intelligence would. So the idea is, hey, what if we have this flexible store of knowledge that understands the relationship between each of our customers? So perplexity, cursor, flock safety as an example. And their end customers, and all the interactions there, in the same way that a human seller would. And so, that data actually is not stored in a relational database, it's stored in much more of sort of a flexible knowledge store that looks a lot like human memory, and it's about updating and cleansing that, whatever it might be, and then when you need to make a decision about this relationship, you can pull on that memory, and you can actually say, hey, how did that last call go? What were the things that stood out? And all those sort of things are stored in one place, and the idea is that, This is an architecture that may scale much better as LMs and tools around LMs get much, much better. And the idea is that we want the product to start to look and feel a lot more like a human in a lot of ways. And so this is a core break from, again, that old paradigm of thinking about things.

Interesting. And then I think in the industry right now, there's talks about this idea of context rot, right? And if the AI has too much information on this particular subject or if you feed it, if the input's too much, then it overloads and starts making a lot more problems or hallucinates more. Does the observation model help counter that at all or is there a moment when there's maybe too much information and there's too many notes on a person or a company that then you would have to almost have the AI prioritize within the doc as well?

I think we're taking a much more opinionated stance on how that all should work because totally, the more you jam into an LLM, you're less likely to maybe get what you want out of it. I do think that there's going to be— I imagine over the course of the next year or two, it's going to be a problem that's going to be worked on pretty heavily, though, is how to retrieve out of these large stores of information the right relevant points. I think there's going to be a bunch of interesting ways of coming at it. I do think it has forced us to be very opinionated though on how we think these things should work and to start to experiment and learn. I think that's sort of the primitive to actually finding things that work. So. Got it.

And then can you talk a little bit about AI kind of buying software as well? Because I think right now I think you're empowering a lot of humans to purchase software to go to market. But do you kind of see a future where AI essentially starts buying as well? And then how does Unify kind of fit into that?

Yeah, it's a really good question. I do think that we're not far away from a world where both buyers and sellers are using agents to run a lot of their day-to-day. And so I think it's a really interesting thing to think about is like, how does the world work in that era? And I do think that buying software is going to change a lot, right? I think we're already starting to see that with very simple transactions, right? Where like, someone instead of, you know, going to an e-commerce store may actually just sit on ChatGPT or Perplexity and do all their research there, right?

Yeah.

We're starting to see like some of the earliest integrations with the LLM providers have been for like some of these shopping use cases. I don't imagine that things are going to change that. I imagine, sorry, that that exact same concept will apply to more transactional B2B software. Yeah.

I mean, I think I remember like Shopify when they came out, Most recently they had a feature on their Sidekick AI assistant where if you say, oh, give me, tell me what apps I should use for reviewing, it'll be a comparison of 3 different Shopify apps that does reviews, then you can choose. But essentially, are you going to be in the top 3 or not starts to be another question.

Totally.

The go-to-market almost dramatically shifts because of this.

It does. It absolutely does. And I think the world's going to change a lot on the transactional side. I do think more traditional sort of B2B selling will continue to stay a lot more insulated from that in the sense that you may do some of your initial research. I think those things will still matter. But I do think that when you're talking about spending $30,000, $40,000, $50,000, $100,000, $200,000 a year on something, all of a sudden you start to look a lot more for, for trust in the relationship and reliability and all these sort of things. Things that are typically human trust and connection matters a lot more for. So I think they'll be a little bit more insulated, but I do see the more transactional side of the market definitely going that direction.

So at least in the short term, we can see a world where you would give an AI agent a budget, like a $1,000 budget, $5,000 budget and say, oh, just pick the best project manager. Exactly. But do you think in the future that that will actually shift towards the larger enterprise software as well? Or do you think that will, let's say, 5 years from now, is that still going to be very, very human touch? Or is that— yeah.

I do think it will still be very human touch to a degree, but I do think a lot of the research and what happens will still happen outside of the buying process. We already see that as being true, but people are going to be reading up on reviews across all sorts of review sites, Reddits, communities, things like that. And I do think that that's going to be done by agents, and that will be a really important piece of the buying process as well.

Got it.

But I do imagine this could be much more of a human in the loop in the actual decision-making process.

Got it. So you said you were an early employee at Ramp. Obviously, Ramp's culture is very heavily talked about as being an incredible one. What kind of learnings have you taken from your time at Ramp and applied it here? And how did you kind of build out the culture here?

Yeah, yeah, totally. I think a lot of how I like to operate and build companies, I think, is very much shaped by my time at Ramp and had an incredible experience there. And so, you know, heavily inspired by it. I think interestingly, one of the most important things is just like a focus on talent and people. I think early on, it's funny, I joined Ramp when the company was, you know, 40 people, and the market at the time was really Brex and Divi were sort of these upstart companies that were much bigger. And I got a lot of really smart people who were like, why would you want to join this company? I had offers from companies like, you know, I had an offer from like Figma as an example, and other like very well-established companies. And the thing that really stood out to me was just the people. And it wasn't that, you know, everyone was like smarter than me or things like that, but it was— there was a certain level of hunger that people had, a certain level of like, I haven't proved myself yet, but I really want to. But the right level of credibility to say like, hey, you know, like these people are legit and I think I would learn a lot from them and really want to work with them. So Ultimately, that motivated my decision to join the company. I think ultimately, at the end of the day, that is a huge piece of it, because a lot of the folks that I, when I joined, were sort of like managers of 1 or 2 people. I ended up, you know, Jeff as an example, who was my boss, Jeff Charles, who's now Chief Product Officer, was the first PM at Ramp and then became the first head of product, and, you know, it's really interesting to look at in retrospect of him becoming the Chief Product Officer of a now $20+ billion company, A lot of these people just sort of stayed and developed, and I think there was a very strong belief in homegrown talent. There was a very strong focus on hiring for slope over intercept, so largely people that were high-growth mindset and could scale into new problems, because at Ramp, the business changed so much every quarter because we were moving so quickly and growing so quickly. We would almost double almost every quarter, and it was just this huge sort of growth trajectory, and so, you know, if you hired someone for one moment in time, they would not scale with the business. So there was that. There was a huge focus on velocity as well, and what that meant in practice. I think people will say velocity a lot, but what does that actually mean? I actually— the way that it was emphasized at Ramp that really resonates with me is it's not about, you know, working 100 hours a week on 100 things. It's actually about being extremely methodical and thoughtful about the 2 or 3 things that you want to get done and holding yourself accountable for getting those things done, and then doing the same thing every single a week, and so that was a big focus. This has become pretty famous at this point, but Ramp would count the number of days since incorporation as just a way of marking the passage of time. We actually do that too at Unitool. So I think those are some big things that took a lot of inspiration from.

Wow. Yeah, and then how did you meet Connor, your co-founder?

Connor and I went to college together, so we know each other from college. Connor actually, funny enough, started a company out of undergrad with one of our close mutual friends named James. And so James is one of my best friends, one of Connor's, and we had— Connor and I had known each other loosely for many years. But when it came time to start the company, James was like, you got to talk to Connor. And the funniest thing is that all my friends would always say, yeah, Connor, like, I always want to try to hang out with him when I'm in San Francisco, but he's always just like working on his like company thing and like trying to hack on some side project.

Yeah.

And I was like, I wanna work with that guy. So that's how we originally got connected.

Do you know why James originally wanted to make the, wanted, like, said like Connor would be like a good fit?

I think we both were set from day one on building a generational company. I think that it was, you know, there's a lot of like, you know, there's a lot of ways to go build a company and, you know, no one right or wrong way to go do it. But I think one of the things that we were really, set on is that we wanted to go build a $10 million plus public company, and that was sort of the outcome that we were shooting for, and we wanted to go after what great looked like there. And I think James knew that we both cared about that thing, and ultimately what that means is that, you know, Connor and I are virtually working all the time. We're extremely passionate about what we do, and, you know, I think it shows up in a bunch of different ways.

When you hire, how do you kind of look for that kind of hunger, and, you know, the slope, right? Like, how do you kind of figure that out?

I look for the things that someone does outside of work, for the most part. I think it's a very good indicator of how they're gonna act in a high-pressure environment. The people that are extremely high slope are, in their spare time, they're trying to learn from people who are the greats at what they do. They're trying to, you know, read books and learn about how things are done. They are maybe working on things on the side. There are things that, even if they're, they don't need to be successful things. They don't need to be things where like they've been, you know, they've created some company and sold to Facebook or whatever it might be. But even just like the willingness and interest in seeing something through and end, I think says a lot about a person. So I look for a lot of those things.

Got it. So, you know, you work at Unify, GTM company. What are, what are your tips on writing a good cold email?

Good question. It's actually surprisingly much, much harder than people think. I would say that first and foremost, you want to appeal to the person on the other side of the email. You don't want to like appeal to the professional or the like, you know, whoever they might be. As an example, when I— the first job that I got out of college, I actually ended with a cold email and the hook in the subject line was just, you know, Rice 2016 grad, like looking for like a summer internship. And Rice is— the reason why that matters is that, you know, I think maybe similar to Papson, but Rice is a small college and the classes aren't that big, but people really look back fondly on the experience and there's a lot of these like sort of like inside stories and jokes. And, you know, I put a couple of those in the note, but I think those things like appeal to the person on the other side and that matters. You're trying to make a connection with somebody and ultimately, like most humans, like appreciate human connection and want more of it in their lives. So I think that's one thing. The other is just to be extremely clear and concise. Those two are actually, I think, tied together in a lot of ways, like you If it takes a lot of words to describe something, you're probably not being that clear on what you're actually trying to communicate. So you'd be surprised, I read a lot of emails where I'm just like, I don't actually know what you want. And I think that's, I wanna know very quickly what someone wants, and if I can be helpful for it, I'm happy to lean in. So I think if you get those two things right, you'll find a lot of success.

Yeah, I mean, this feels like a very underappreciated skill set to have, right?

Totally.

And obviously, if you have run a B2B, business, you can use Unify and automate part of that. But if you're just an individual going around, but there's so many doors you can open just through cold DMing people, cold emailing people. I definitely know that just because I've interviewed a bunch of folks and a lot of those were through cold emails. So another thing I wanted to ask is you posted, I think it was on X, about this O3 prompt that Toby from Shopify, I think you were inspired by him. And the prompt went something like, tell me something special or unique you've noticed about me, but you think I haven't realized about myself yet. So you would ask this to O3. What learnings did you get from asking this and how has it helped you in changing your behavior?

Yeah, well, one meta-learning is that AI knows you much better than you think it does. And it's actually, it's wild how well it does. I actually, I ran that one again recently and it gave me a new updated answer. It was crazy compelling. I think that there's so many ways to build a company and to be a founder that, you know, sometimes you get a little bit lost in thinking about like, well, this other person is doing this thing, or, you know, and they're trying it in this way, and maybe I should be doing it that way, even though it doesn't feel like congruous with who I am or how I like to do things. And I think the biggest thing that this prompt really told me is like, it's okay to like be yourself and really lean into your strengths, right? And I think that, you know, my ChatGPT experience knows a lot about me at this point and it sort of said some form of that. But I think that's a really like important thing to keep in mind is that there's, there's no single way to do things in the right way. Of course, you know, you can borrow a lot of things that you learn along the way, but you should also have some version of like, hey, what resonates with you and What is your superpower that you want to lean into? And so it's a good reminder of that.

That makes sense. So I think you, you've previously mentioned about this, this idea of the outbound quarterback. What is the outbound quarterback? And what are examples of different kind of plays that they think about?

So the outbound quarterback is this new role that's emerging that's incredibly valuable for companies. It's this idea that you can have one person who can actually scale up activities that look like an army of sellers, but have it be one person behind the scenes who's sort of orchestrating the system. And these people tend to be quite different from your traditional go-to-market folks. Sometimes they come from technical backgrounds, like they've been an engineer before. Sometimes they come from more of like a professional services background, like a consulting or banking type background. I think those profiles are often actually underrated in this context. And it's somebody who, is experimental and systems-oriented, so you can actually give them ideas, they can know how to test through ideas, try them out with rigor, see what works, see what doesn't, put things in place that do work. And so we find that those types of people that today wear titles like, you know, they can wear a revenue operations title, they can run a business operations title, they might come from a sales team and actually be just like a more technical person there. They're incredibly valuable for companies. And we're seeing the demand for this talent go through the roof with like very, very high salaries, very compelling offers, and Most, most of the top companies today are asking me for these profiles of people, and they're just very hard to come by.

Got it. Yeah. Do you think an AI can, can replace at least part of the work?

I do think so. Yeah, I do think an AI is and will replace like quite a bit of that work. But I do think there's always an element of when you have a piece of software that makes your life easier, humans are incredibly gifted at being creative and thinking outside the box. And I think that's going to open up a ton more bandwidth to do those things, which will make it an even better combination.

Yeah. So you were recently— when you raised recently, you were on TVPN. And then afterwards, you sent a thank you note through X, through a video via Russ Hanneman.

And what—

yeah, talk about what went through your head when you wanted to— Was that originally the thought beforehand, or like, yeah, how did that come about?

Yeah, we, so we actually, as part of a batch, we actually created, so the campaign we ran is we created about 25 Cameos for select partners and customers and folks that we know. And the idea was just like, yeah, how do we do something that's truly special and interesting, right? As like a thank you to a lot of these folks who have been really instrumental in helping us be where we are today. And so we hadn't seen this one done before, but we thought how cool it would be if you could get like a personalized video as a thank you from somebody that you really think is cool and interesting. And so for the TVPN guys, who are awesome, we created this one from Russ Hanneman. And it was a ton of fun to do that.

Yeah, and he's always great. He's always on character. He always knows what to say. And this is, I guess, another kind of thing you talked about with the marketing, the distribution side, right? It's like obviously you're gearing that towards your existing customers, but people will see it and people will share it as well, right?

Totally.

Makes sense. Yeah. And then in terms of some quickfire stuff, what books do you keep coming back to? Do you have any go-to books?

Good question. I'm not I'm not somebody who I would say comes back to books super often, but I do try to read quite a bit just to learn new things and absorb new information. Right now I'm reading Marc Benioff's book about the story of Salesforce. Actually, tons of good nuggets on marketing that I think are still very relevant.

He's an insanely good marketer.

He's an insanely good marketer, yes. And I also do really appreciate history more broadly because I do think a lot of things sort of tend to repeat themselves. Recently read all of Ray Dalio's books as well, just like sort of understanding economic cycles and things like that. So I'm trying to think if there's any one too that like stood out recently for me. I did, I did read, which one was it, escaping me now. Those are some good ones. I'm trying to think, there was one more that I wanted to mention, but forgot it.

Yeah, it happens, it happens. What keeps you up at night right now? Whether it's a good thing or a bad thing?

I'm just incredibly excited to be building right now, building a company. I think when we started in 2023, it was really obvious to us that the world was going to change because of what was happening with AI. And I think Connor and I were lucky that we were earlier to things to understand, hey, these models are incredibly creative, are going to change how knowledge work knows as we see it today, or And I do think that it is a time where success feels abundant. It feels like there's so much opportunity for those who are excited and eager to go out and get it. And I think that's incredibly exciting, and I'm just so grateful to just wake up every day and think about how the world could be different, put into play. And it's a little bit of a, it's also scary in the sense that no advantage is a big advantage. Anybody can catch up quickly, and so the only thing you can do that compounds is just building a company that has a culture of just continuing to innovate it. So it's a lot of fun.

Amazing. Well, thank you. Thank you for your time, Austin, and hopefully we'll, you know, you'll start expanding out into Japan as well.

Yes, would love that.

Amazing. Thank you.

Thanks for having me on the show.

Thank you.