Why the Sequoia thesis changes the unit economics of every services vertical, and how FormBridge is applying it to global company formation and back-office work.
"The next $1T company will be a software company masquerading as a services firm." — Sequoia Capital, Services: The New Software
That sentence reframes how to build any venture-scale business in a services vertical. This post is about why it's right, and how FormBridge is applying it to global company formation and back-office work.
For every dollar spent on software, six dollars are spent on services. SaaS companies compete for the $1. Services firms compete for the $6. Historically the $6 budget was untouchable by software — human judgment, variable work, relationship-dependent outcomes. The whole point of buying a service was that you weren't trying to run it yourself.
AI didn't just make software cheaper. It made outcomes cheaper. The moment an agent can file your taxes — not just help your accountant file them — the customer has no reason to keep writing two checks. The $1 and the $6 collapse into a single, much larger contract.
That's the structural shift. Software companies can now eat the services budget.
The Sequoia framing distinguishes two kinds of work:
Modern LLMs passed the threshold on intelligence work. They haven't on judgment work. That's a specific, testable claim — not a vibes statement.
The right services businesses to build as Services-as-Software are the ones heavy in intelligence work. Company formation, EIN applications, bookkeeping, tax preparation, compliance calendaring — all of these are overwhelmingly rule-based. Competitive advantage flips from who has the best people to who has the best prompts, tools, and evals.
The 2025 wave of AI-assisted professional services ("Harvey for lawyers", "Cursor for engineers") is a copilot wave. A copilot sells productivity to the professional; the professional still charges the client.
The copilot captures the $1 (a SaaS license). The $6 still goes to the professional's firm.
An autopilot goes direct to the end customer. No professional intermediary. The autopilot charges for the outcome, captures the $7, and sets prices against a competitor (the professional firm) that has to pay salaries.
Copilot businesses are SaaS businesses. Autopilot businesses are services businesses, priced like SaaS, with gross margins that compound as prompts and models get better. That's the arbitrage.
The formation-and-back-office vertical is a good test of the thesis for specific reasons:
At FormBridge we have a specific internal definition. An agent doesn't count as an autopilot unless it meets six tests:
A chatbot that helps you think about filing your taxes is a copilot. An agent that files your taxes, with a CPA signing the return before it leaves the platform, is an autopilot.
The distinction is a product decision and an economic one. Copilot companies are valued as SaaS. Autopilot companies capture the full work budget — which is where the $1T outcomes come from.
FormBridge ships eight autopilots against the full back-office surface:
Each one is built against the six tests above. Each one has published cost budgets. Each one ships with evaluation cassettes covering the happy path plus ≥20 edge cases.
The claim isn't that AI is magic. The claim is that formation and back-office services are dense in intelligence work, that modern models handle intelligence work, and that an autopilot priced against human-delivered services has structurally better unit economics.
Sequoia's estimate — the next $1T company here — may be optimistic, may not. What matters for FormBridge is narrower: the window for a pure-play autopilot to out-execute incumbents is roughly 12–18 months. Every existing services firm is a copilot at best today. The ones with the biggest customer bases have the biggest innovator's dilemma (turning your $6 services revenue into $7 autopilot revenue means firing the people who delivered the $6).
That's the window. That's why we're building fast.