Running a business7 min read

We Gave Away a Detailed Scope for Free. Here’s Why That’s the New Normal.

AI collapsed the cost of detailed scoping — so we give ten-page scopes away before a contract is signed. What changed, what got worse, and how to evaluate a proposal in 2026.

A founder came to us recently with an idea for a two-sided staffing marketplace. She knew her industry cold — she’d worked in it for years, she was an active user of the platforms she wanted to compete with, and she could tell you exactly what those platforms got wrong. What she didn’t know was what she was actually asking for when she said “I want to build a website where offices can book professionals.”

That sentence sounds like a website. It’s not. It’s inverted booking logic, availability matching, post-completion billing, license verification, identity checks, bidirectional reviews, worker classification questions, payment processing, dispute handling, and an admin layer to keep all of it honest. None of that was visible to her, because none of it is visible to anyone until somebody maps it out.

So we mapped it out. Before any contract was signed, we sent her a ten-page scope document: five phases, detailed feature breakdowns, a table of everything that wasn’t included, assumptions, client responsibilities, timeline, and pricing ranges. For free.

Five years ago that would have been a terrible business decision. Today it’s just how we work, and the reason is AI. Here’s what’s changed, what hasn’t, and what you should be looking for whether you’re buying development work or selling it.

What AI actually changed about scoping

It changed the cost of detail.

A document like the one we sent used to take a week of senior time to produce — which is exactly why most agencies wouldn’t produce one until you’d signed something. Pre-sales effort had to be rationed. The result was an industry standard of vague two-page proposals (“Phase 1: Discovery. Phase 2: Build. Total: $80,000.”) that protected the agency’s time at the expense of the client’s understanding.

With AI in the workflow, the mechanical parts of scoping — structuring the document, enumerating standard features of a known pattern, drafting assumptions and exclusions — collapse from days to hours. The expensive part is no longer the typing. It’s the judgment: knowing that a staffing marketplace has inverted booking logic, that worker classification is a legal landmine the client needs their own counsel for, that the payment flow is the highest-risk part of the build and should be priced and phased accordingly. AI doesn’t know your client’s business is non-standard. You do, or you don’t.

So the economics flipped. Detailed scoping pre-contract is now cheap enough to give away, and giving it away is the single best trust-builder available — because the client walks away understanding their own project better than they did before they talked to you, whether or not they hire you.

What AI made worse

Here’s the uncomfortable flip side: every agency can now produce a beautiful, detailed, ten-page proposal. Including the ones that have no idea what they’re doing.

A polished document used to be a weak signal of competence — at minimum, someone spent real time on it. That signal is dead. An AI can generate a plausible-looking scope for a “marketplace platform” in minutes, and it will look great, and it will quietly miss everything that makes your marketplace hard. Plausible and correct are different things, and the gap between them is where projects die.

Which means if you’re the buyer, you have to evaluate proposals differently than you did three years ago. The questions that matter now:

Does the scope name what’s specifically hard about your project?

A generic scope describes features. A real one describes risk. If the document for a marketplace doesn’t mention payment flow complexity, verification workflows, or regulatory exposure, it was written by someone — or something — that doesn’t understand the domain.

Is there a “What’s Not Included” section, and is it specific?

This is the most honest section of any scope document, and AI-generated boilerplate almost never gets it right, because exclusions require knowing where projects like this actually sprawl. Native apps, enterprise dashboards, third-party integrations, content creation, legal review — every one of those is a future conversation, and a scope that pretends otherwise is setting up a fight.

Is discovery a paid, client-owned deliverable?

If discovery is “free” and bundled into the build, you own nothing if you walk away. We structure it the opposite way: a fixed-price discovery phase whose output — requirements, data model, wireframes, technical spec — is yours upon payment, detailed enough that any qualified team could build from it. Including someone who isn’t us. That’s not generosity; it’s the only structure where the client isn’t hostage to the agency.

Who actually does the work, and will they say so?

Our client asked us directly, and the answer was: mostly me, with specific things farmed to specific people I’ll name. AI has made it trivially easy to look like a twenty-person shop. Ask, and watch how the answer is given.

What didn’t change at all

The client’s real question was never about the timeline. We asked her what mattered most, and her answer was the one we hear over and over: if something is going to take longer, just tell me.

No amount of AI changes this. Software estimates are probabilistic, third-party integrations slip, and somewhere in month four something you didn’t plan for will show up. I know because it’s happened to me — I’ve delivered a project months late, and the thing that saved the relationship wasn’t the eventual delivery. It was that the client never had to ask where things stood.

So the part of a scope document I’d weight most heavily in 2026 is the part most proposals still don’t have: a communication commitment. A written cadence, a promise that risk gets flagged the moment it’s known rather than at the deadline, and timeline confidence stated honestly per phase — “fixed” for discovery, “range, firmed up after discovery” for the build. A roadmap with honest confidence levels beats fake-precise dates every time, because the precise ones are fiction and everyone signing the document knows it.

How we’d tell you to scope a project right now

If you’re a founder buying development work: don’t accept a two-page proposal for a six-figure build, and don’t be impressed by a forty-page one either. Look for the document that makes your project more legible to you — that names the hard parts, draws the boundaries, ties payments to milestones you can verify, and tells you who you’ll actually be talking to. We’ve written before about why we quote fixed fees and what flat-rate pricing actually buys you — the same honesty test applies to scopes. Get competing quotes; any agency confident in its scope will encourage it.

If you’re an agency: stop rationing your thinking. The detailed scope you used to guard behind a signature is now your best marketing, your best qualification filter, and your best protection against scope disputes later — and AI has made it cheap enough to produce for every serious prospect. Just don’t let the tool do the judgment part. The document is free. The twelve years of knowing what belongs in it are what the client is paying for.

The founder with the marketplace? She read the scope and told us it was exactly the vision she had in her head — which is the entire point. She didn’t know what she was getting into. Now she does, before a dollar has changed hands. That’s what scoping is for. And if you’ve got the next one, you know where to start.

The newsletter

Stop firefighting, one email at a time.

Plain-English notes on keeping your platform fast, stable, and out of your way. No fluff, ~2x a month — join 2,500+ founders.