170. 75% of SaaS Companies Will Disappear ft. Brett Queener (Bonfire Ventures)

And, you know, I think this is the biggest prediction that I had made 18 months ago, which is the return of the buy versus build dilemma. And it’s very real. And so I think good news is I think it eliminates a lot of frauds. I think a lot of software companies that don’t deserve to just go away.

Josh Schachter [Host]:
You’re listening to Unchurned, brought to you by the Gainsight Podcast Network. 18 months ago, Brett Queener predicted that 75% of customer-facing software companies would disappear. So he did something unusual. He turned his own AI agents loose to grade his predictions. And the verdict? He was right. The moats that software companies counted on are now just the price of entry. But here’s what he found next. In an agentic world, you don’t buy software anymore, you hire it.

Josh Schachter [Host]:
And the only thing that separates winners from the dead, something he calls the judgment layer. I’m Josh Schachter, and today on Unchurned, Brett Queener, managing director of Bonfire Ventures and early Salesforce executive, on why judgment may be the only moat left. All right, Brett, we’re live here.

Brett Queener:
Hey Josh, how are you, man?

Josh Schachter [Host]:
I’m doing well. I’m doing well. You just published your next chapter of movie.

Brett Queener:
You remind me you’ll get, you’ll finish by 2027 when it’ll be completely obsolete because it’s too long. That one.

Josh Schachter [Host]:
Yeah. Is that your thing? Is that you like write War and Peace so that by the time people are done, like, you know, they don’t catch the ending, you’re onto the next one?

Brett Queener:
I like to think that it is long, but it is interesting enough that you stay with it and you will see in my writing, there is, as any good narrative arc, there’s a setup, there’s a buildup, and then there’s somewhat of a happy ending. And so sometimes I bury the lead, et cetera, but I, it’s important as we know with AI and the rest of it to give people proper context. This was just a little long because we’re at another moment. You know what I think for the last 2, 2 and a half years, we’ve been in the GPT moment. And all the startups and the thing, da da da da da, and like these winners and Cursor and all this, the GPT moment. It’s like the amazing copilot. But the copilot is still very much, I’m a human and you have an assistant and you help me do my job. But I am the OS, I instruct.

Brett Queener:
The last 60 days were in a brand new era, what I’ll call the cloud code, cloud co-work. And the reason I call it co-work versus code is that code seems, oh, I don’t know how to write code, co-work. Allows you to get in initially and then you’ll get into cloud code because you’ll be like, oh boy, or the Codex and then these new models. And the interesting things about these new models that GPT and others are building is, uh, they’re using their own code building agents to build the next versions of their models. Like, and what we’re moving to now is that, um, agents are like employees., they don’t have AGI, but they’re semi-autonomous, but you don’t give it tasks to help you do a job. You assign it an outcome. It can be a big outcome. And then you tell it to build the team to work with you to go do the work.

Brett Queener:
And instead of you instructing every step of the way in the old way, you’d be in a prompt and then you’d keep instructing and then, oh my God, the prompt got too big and it got memory, like that whole thing. Your role as a human moves from like, I’m helping, I’m giving instruction in doing the job, but it’s helping me do it to you’re not doing the job. Others are doing the job and you’re moving to review and management. And that’s the new era that we’re in.

Josh Schachter [Host]:
And you were telling me a second ago off the air that I got.

Brett Queener:
To start building out. Yeah. And so like the point of the matter is I can tell you it, you can read the Slack and you can read the stuff, but can’t I.

Josh Schachter [Host]:
Wait 6 months for the next evolution of it when it’s done?

Brett Queener:
No, because If you’re in technology, whether you’re an investor in technology or whether you work in technology, um, it’s very important for you. It’s not like, oh, if you don’t do it now, you’re obsolete, but you will be obsolete at some point for you to grok and understand what this is. We just talked about the new responsibilities you’re working on at Gainsight, you know, and at the end of the day, you can’t properly approach those responsibilities because in your responsibilities, you’ve got, okay, What do we, you know, what do we, what’s the issue? What are we trying to accomplish? Where do we stand? What are the metrics? And you’d be like, all right, well then, okay. And then what are the people? What’s the team and how do we organize and specialize? And I got to go manage the team. And then there’s the information loops, blah, blah, blah, blah, blah, blah, blah. Or you’re trying to build a product, whatever it is for a customer. You need to go into the products, these AI products, and go deep just a little bit. So then you go, oh boy, let me rethink this.

Brett Queener:
What’s possible? What’s possible.

Josh Schachter [Host]:
Now?

Brett Queener:
Um, because, and then it almost becomes irresponsible for you not to do that. And when I say get into the tool, what I just mean, like, look, Claude code is scary to you. Go to, go, you know, look, start with vibe coding, vibe code an app. You’re like, oh, this is pretty cool. And in the vibe coding, what you’ll understand is, oh, the old world of, like, I think we should solve this problem. I don’t know what it looks at. Let me go write a bunch of PRDs and then hand it to dev. And then they build stuff and then it comes back and there’s information loops, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah, blah.

Brett Queener:
And it doesn’t work. So God forbid you need to deploy gainsight because your customer success. People have to help people figure this out. Just the. Oh, wait a minute, I see something visual. And this is the part of the article on the judgment layer I didn’t mention. This is the revenge of like, you know, all the people that were in the humanities and the classics were shutting down those programs across the colleges to pump out STEM kids. It’s like, uh, it’s like not revenge of the nerds.

Brett Queener:
It’s like revenge. It’s revenge of the dilettantes. Like if you have good judgment, whoo-wee. And so it was just this whole back and forth, like, no, no, no, just do this. Instant feedback. Oh, we’ll just do this. Then you’re like, oh my God, I just came in here to build some simple app and oh my God, let’s just keep going. Right? I did this yesterday.

Josh Schachter [Host]:
What was your article on? We didn’t actually introduce your article to the audience when we pressed record.

Brett Queener:
No.

Josh Schachter [Host]:
But who are you actually? We haven’t introduced that either.

Brett Queener:
We’ve done this before. I don’t know who I am.

Josh Schachter [Host]:
Yeah, but it’s not like my, it’s not like the listeners are like waiting with bated breath for like the quarterly Brett episode.

Brett Queener:
So I am a venture capitalist, but I started my career in software 30 years ago at the beginning of moving from, uh, VAX to client-server with email. Came out of graduate school at the beginning of the internet, uh, was an early executive at salesforce.com for decade plus, ran a co— co-ran a startup, which is a successful outcome. And I’ve been an investor in seed stage companies. And I’ve been writing for the last 16 to 18 months for my own therapy. Josh says the pieces are long, but you know, my therapy sessions are long. And by talking and thinking and writing, it forces sort of a clarity of thought that is therapeutic on what this all means for software industry.

Josh Schachter [Host]:
So Tales from the Bonfire, my favorite SubSec blog channel other than my own Unchurned Substack channel. Talk about the moats.

Brett Queener:
Talk about the moats. Well, the image in the latest post has a picture of me as a medieval knight in a Patagonia vest, because VC, I don’t have a Patagonia vest, chasing this moat that keeps running away from me. And what I’ve been trying to establish With the onslaught of AI, if you’re a founder in a software company or investor, what are you investing in that’s defensible as the market moves so fast and technology moves so fast? And so I had a new thought around what I think a new moat or defensibility is, but the latest article, first I had AI and a bunch of agents go out and tell me, should I stop writing? ‘Cause I’ve been writing about this and I thought I was prescient. Like I wrote 18 months ago that 75% of all customer-facing software companies would disappear and 50% of all categories would disappear. At the same time, we’re raising a fund where we invest in application software companies. Um, ah, but we’ll pick the right ones.

Josh Schachter [Host]:
That must feel.

Brett Queener:
Good. Um, and what was interesting was would I, and so I wanted to analyze, was I right enough that I deserve the right to write this piece or should I just hang it up? It’s not that useful. So the first half of the piece really goes and just reflects on the journey of the last 18 months around all the moats that we thought were true that have sort of disappeared or now are just a right to play, but not a right to win. They’re just table stakes. So you don’t have them, you can’t play., but they’re not a right to win. Um, which I think is helpful.

Josh Schachter [Host]:
Can you have a right to play without it being a right to win? Don’t you need a right to win, especially if you’re in your— No, a.

Brett Queener:
Right to play means you have a right to play the game. Like, you know, if you’re in basketball, you have two legs, you can move, you can breathe, whatever. Like you’re somewhat fit. That’s a right to play a game. It’s not the right to win the game. And so what I thought before were rights to win. I.e., a moat. Like if you got this right, you could win a game.

Brett Queener:
Now I’ve been like, no, this is.

Josh Schachter [Host]:
The price of entry. The system of record, all the context.

Brett Queener:
Yeah. And so, and so this, I think, and reason I did that first, I also, it’s part of the journey. I think everybody’s been feeling this. They’ve been chasing the thing. They thought they were doing something and then they’re like, oh my God, some of my founders have to find and refine product market fit every 3 to 6 months. That wasn’t the case 5 years ago. And then the agents work with me on this. Yeah, you have a right.

Brett Queener:
You’re pretty good, right? So let’s write your next conclusion. So the next conclusion really was around what I call the judgment layer, which is in the end, as we think about moving agentic, What is the judgment layer of your product? Meaning, if you’re bringing a product to market and it has enough context such that somebody just wouldn’t go build it themselves, does it ship with judgment? Does it know enough about the skills that I want to get done if I’m moving in an agentic world that you’re instructing agents to go do that I should just, I should use yours versus my own. That’s the baseline. But the big one is really the level 2. And this is something was never the case in SaaS or enterprise software, which is what is the judgment that an organization or person has, how they want to work or operate, and how do you capture that and feed that back to the user? Because that’s the marvel about using these agentic products. So whether it’s you’ve built a lovable app and you’re giving it feedback, or as I was writing this piece with the agents and they self-organized which agents would go do, and then they actually inserted a skeptic agent because they were getting too sympathetic. They were liking me too much. And I was like, stop it.

Brett Queener:
Um, when you give them feedback, they improve so quickly. And that memory stays there, that you’re much more willing to dig in and give it feedback and give it feedback and give it feedback and give it feedback. Because you said, Brett, this is pretty interesting. Chuck said there were some em dashes in the article, but you said there wasn’t a lot of slop. Because I went in, it produced, I went in each, and then I had a co-writer with me and an editor. We fleshed it out, but I trusted it.

Josh Schachter [Host]:
And like you would think— By the way, Brett’s talking about his most recent article. What was the title of your article, by.

Brett Queener:
The way? Whatever. In the end, I think judgment may matter the most.

Josh Schachter [Host]:
Something like that. Okay, cool. On his Tales from the Thought Factory. But anyway, you keep on saying things that the audience doesn’t have context for. They weren’t in the text thread, right? So I was applauding you on having.

Brett Queener:
No AI slop, like your article. And then Chuck was like, there’s some em dashes in it, because AI does like em dashes.

Josh Schachter [Host]:
Chuck hates em dashes. If you ever want to piss Chuck off like I piss you off, just give him some em dashes.

Brett Queener:
It’s fine. But he hates em dashes because it’s usually associated with slop. Right. So, and if you just think about this, if you’re building product, if you’re using product in the past, you use some product like, ah, this sucks. You’re like, oh, you got to go in and configure your personal settings or your configuration or change a bunch of workflow rules. Like, and what, that gets logged as a request or CSM hears it and maybe it shows up in a product 9 months later. And I’m already like, eh, but now I’m like, dude, this is wrong. Go do this.

Brett Queener:
Okay, thanks. We’ll keep that in note.

Josh Schachter [Host]:
We need to buy software anymore.

Brett Queener:
That second layer is very important for a software provider. Next question. You said, do I need to buy software anymore? Yeah. If it’s a given problem that you’re trying to solve, That is very easy to make sense of the context and it’s somewhat generic and you don’t actually care about best practices imbued in it. Like you think you know better how to do something. No, but if it’s, I don’t know, I’m a PI lawyer and I’ve got a team of 250 people who only work with PI lawyers and understand how PI lawyers work and they give me something out of the box where all 80% of the hardware work is already there and it gets better because it works with all these PI lawyers. So I, you know, so it’s injecting the best practices in an agentic way across how PI law is done and it gets better over time. And then it allows me to inject the way I want to work, the way I write, the way I look at cases.

Brett Queener:
Yeah, I’m going to buy that software. And the way I think about it is stop talking about buying software. Do you want to, you’re hiring an employee. And so the question is, do you.

Josh Schachter [Host]:
Want to hire an engineer?

Brett Queener:
Yeah. But it’s an interesting question. And, you know, I think this is the biggest prediction that I’d made 18 months ago, which was the return of the buy versus build dilemma. And it’s very real. And so I think good news is I think it eliminates a lot of frauds. I think a lot of software companies that don’t deserve to just go away.

Josh Schachter [Host]:
I hope, by the way, are you putting these predictions on Kelsey? Because you could be making a lot of money with all of your AI predictions.

Brett Queener:
No, no, it’s not a bad idea, but no.

Josh Schachter [Host]:
Until next.

Brett Queener:
Time. Thank you.