I've been running screening calls for forward deployed engineers at Mintlify, and I keep having the same conversation. An ex-founder, two or three years into a company that didn't work out, wrapping things up, looking for something stable. I recognize the tone of voice because it was mine a year ago. One candidate put it as "only founders get me," and yeah, same.
That pipeline is full of founders on purpose. Candidates keep asking me two versions of the same question: why does a docs company need forward deployed engineers, and why founders specifically? Both answers come from one belief about where SaaS value is going, so I'm writing it down.
The cost of product is going to zero
I ran Helicone from first commit to acquisition, and the difference between building software in 2022 and building software now is not subtle. A founder I interviewed last week told me his two-person team shipped in one year what would have taken three before coding agents. I believe him, because I spend my evenings vibe-coding iPhone games, and the games actually ship.
When anyone can stand up a credible clone of your feature set in a quarter, the feature set stops being a moat. There are exceptions, the SolidWorks of the world, where twenty PhDs spent decades baking in secret sauce that no agent is going to regenerate. If that's your product, congratulations. For everyone else, "we have the best product" is a claim with a shrinking half-life.
So where does the value go? I think it consolidates into three pillars: security, services, and infrastructure. If your company doesn't sit on at least one of them, the moat conversation gets uncomfortable within five or ten years.
Security first. Enterprises don't buy features, they buy confidence that nothing bad happens when those features touch their data. Compliance, audits, tenancy guarantees, surviving a hostile review from a security team at a company a thousand times your size. None of that can be vibe coded, because almost none of it is code. It's accumulated trust, and trust compounds slower than software. It also appreciates as products get cheaper: when every demo looks the same, procurement stops evaluating demos and starts evaluating blast radius.
Services used to be where margins went to die, which is why every SaaS playbook of the last decade said to run from them. A $500K self-hosting contract didn't justify parking an engineer inside a customer for months; you either walked away from the deal or handed the pain to a systems integrator and lost the relationship. AI flipped that math. The expensive part of forward deployment was always debugging someone else's world: their homegrown git system with its own event bus, their Kubernetes setup, the internal service nobody documented. An agent can grok that world in an afternoon. Contracts that were money losers two years ago are profitable today, and the work compounds, because the adapter you build for one enterprise's weird git setup is the starting point for the next enterprise's weird git setup.
Infrastructure is the third pillar and the quietest one. It means being the thing that is running, in their cloud, wired into their systems, on the critical path of how their teams work. A tool gets replaced the moment a cheaper clone shows up. A deployment that a hundred internal teams depend on does not. Self-hosting used to be a concession vendors made through gritted teeth. I've come around to thinking it is the moat.
You can argue for other pillars. Distribution is the strongest candidate, and it's real, but I'd say distribution gets you into the building. These three are why nobody can throw you out of it.
At Mintlify we're betting mainly on services, with a growing position in infrastructure. Not because a strategy deck said to, but because customers with very large budgets keep pulling us in that direction, and when someone is willing to pay millions of dollars for a thing, the correct response is to listen.
The founder job, minus the fundraising
Which brings me to the second question: why staff this with founders?
Because the job is the founder job with the fundraising removed. An FDE walks into an enterprise that bought a product which doesn't quite work inside their walls yet, and closes the gap. The way I describe it on calls is "take in pain, excrete product." Nobody has asked me to stop saying it, so it stays.
There is no spec. There's a customer with a workflow, a pile of internal systems you've never seen, and a contract that assumes somebody figures it out. Founders already have the scar tissue for this. They've sat in a coffee shop next to their first customer, watched them click through the broken flow, and shipped the fix that night. They know the difference between what a customer says and what a customer needs, because they paid for confusing the two with their own equity.
Enterprises don't buy the product you have today. They buy the version of it that works at their scale, against their systems, and somebody has to go zero to one inside those walls. That somebody is, functionally, a founder with a salary.
What should I do?
If you're an engineer feeling like you spent years learning things an agent now does in seconds, you're not alone. I hear it on almost every call, usually right after the candidate tells me how fast the new tools have made them, and right before they wonder out loud what that means for their career.
In high school I had a math teacher who told me something that stuck with me forever. Before a quiz I asked him how much of it was integrals versus derivatives, and how many I needed to get right for an A. It was calc 1. He said that was the wrong question: just learn the material, and the grade will follow. Then he made a point of repeating it to me roughly once a week. Come learn the material. Stop asking about the grade. If you index on the material, everything else follows.
The frameworks, the syntax, the accumulated trivia of your stack: that was always the grade, and the grade is being repriced. The material was never the syntax. The material is walking into a system you don't understand, next to a person who is in pain, and leaving both better than you found them. Software happens to be the best problem-solving training program ever built, and you've been enrolled for years. That training is worth more now, because one person who can actually diagnose a problem now ships what used to take a team.
So keep solving problems. The rest will follow.
(And if closing the gap between a product and a Fortune 50's walls sounds like your kind of problem, I'm hiring.)