The AI consultancy proposal that undersells scope on purpose

A competitive AI consultancy pitch creates real pressure to promise a broad, ambitious scope, because a narrower proposal can look less impressive next to a competitor's more expansive one. That pressure produces proposals that win the pitch and then struggle to deliver, because the scope was sized to impress rather than to actually be achievable in the stated timeline.

The proposals that hold up under delivery deliberately undersell scope relative to what the client might want, and are explicit about why: a narrow, well-defined first phase that can be delivered with confidence, with a clearly stated set of follow-on phases that depend on what's actually learned during the first one, rather than promised upfront as if the unknowns were already known.

This is a harder pitch to make sound exciting, and it's also the version that produces a client who trusts the next recommendation, because the first one was accurate about what it could deliver rather than optimistic about what might be possible under ideal conditions that rarely hold.

Undersizing scope on purpose costs some pitches. It's a reasonable trade against winning pitches with promises that don't survive contact with the client's actual data, actual team capacity, and actual organizational friction, all of which are easier to see honestly from outside the sales conversation than from inside it.