How to Scope an AI Agent Project Before You Pay for Anything

A scoping call should happen before any contract gets signed, and it should cost you nothing. If an agency wants payment before they'll even talk through what you need, that's a sign to keep looking. A real scoping conversation is where most of the risk in a project gets removed — not after the build starts.

What it should actually cover

• The one task, described exactly. Not “automate our sales process” — something like “score inbound leads from our website form and add them to HubSpot with a priority tag.” If it can't be written in one sentence, it isn't scoped yet.
• Which systems it touches. Every tool the agent needs to read from or write to should be named, along with who has admin access to set up the connection.
• What counts as done. A specific, checkable outcome — “every new lead gets scored within 60 seconds” — not a vague feeling of “this seems to be working.”
• What's explicitly not included. This is the line most projects skip, and it's the one that prevents scope creep later. Write down the related things you're choosing not to build yet.

Two questions worth asking the agency directly

“What happens if this doesn't work the way we expect?” and “Who owns the code when this is done?” The answers tell you a lot. A good agency has a clear, rehearsed answer to both. A vague answer to either one is worth pausing on.

Why this step gets skipped

It's tempting to jump straight to building, especially when the problem feels urgent. But an hour spent scoping properly routinely saves weeks later, because it's where misunderstandings get caught while they're still cheap to fix — a sentence in an email instead of a rebuilt feature three weeks in.