By the time an RFP reaches the street, most AEC firms are already behind. Capture-management research consistently finds that 40 to 80 percent of buyers have a preferred firm before proposals are ever submitted. The proposal does not decide the pursuit. It usually ratifies a decision the client started making months earlier.
Reacting to the RFP Is Already a Losing Position
The RFP is a public document. Every shortlisted competitor gets it on the same day, reads the same scope, and answers the same questions. If your knowledge of the client begins there, you're competing on writing speed and page layout — and neither of those wins work.
The firms that win consistently do not treat the RFP as the starting gun. They treat it as the midpoint. By the time it drops, they already know what the client cares about, who is evaluating, and where the incumbent is weak. The document just tells them the format.
This is also why a go/no-go decision made with zero client intelligence is a coin flip. You cannot honestly score whether you understand the client's real priorities if nobody did the research. The firm ends up chasing pursuits on gut feel and submitting into selections it was never positioned to win.
What Client Research Actually Means in AEC
"Client research" gets dismissed as relationship-building — golf, lunches, conference small talk. That is part of it, but it is not the useful part. Useful client research is specific, and most of it is available before anyone picks up a phone.
| Question | The reactive firm | The firm doing client research |
|---|---|---|
| Who is funding this? | "It's in the budget" | The exact source and its calendar — an IIJA-backed program with a hard deadline, a bond, a CIP line |
| What does the client care about now? | Whatever the RFP says | What the agency has said in board meetings, capital plans, and public updates over the last year |
| Who is evaluating? | "A selection committee" | Whether it's engineers, elected officials, or end users — because each one reads a proposal differently |
| Who holds this work now? | Not sure | The incumbent, their performance, and where they are vulnerable |
| What went wrong last time? | No idea | The specific complaint — the schedule slip, the change orders, the communication gap — this procurement is meant to fix |
None of that requires inside access. It requires someone to do the work before the RFP forces a scramble.
The Intel Is Worthless If You Can't Act on It
Here is the part that connects client research to the rest of proposal operations, and the part most business development conversations skip.
Good client research produces a specific instruction: lead with these three projects, put these four people forward, build the narrative around the client's real hot button. That instruction is the entire payoff of the research.
Then it lands on a proposal team that has to actually find those three projects and reformat those four resumes — and that is where it dies. If pulling and tailoring the right content takes a week, the team runs out of runway and ships the generic version: the standard resumes, the closest-match project sheets, the boilerplate narrative. The research was done. It just never made it into the document.
This is the capacity myth again. A firm can have the best client intelligence in the market and still submit as one of the pack, because the gap between knowing what to feature and being able to produce it is too wide. The intel and the retrieval layer are two halves of the same problem.
RFPM.ai exists to close that gap. When staff and project data is structured, "feature these three projects and these four people" becomes a filter and a format choice — not a week of rebuilding.
Where to Start
You don't need a dedicated capture manager to do this well. You need a habit.
- Run real debriefs. Win or lose, capture why — in writing, within a week. Your own past pursuits are the cheapest client intelligence you will ever get.
- Build a one-page client intel sheet. Funding source, current priorities, evaluators, incumbent, known hot buttons. Require it before the go/no-go meeting, not after the RFP drops.
- Move the go/no-go decision earlier. If you're deciding when the RFP hits the street, you're deciding too late to do anything with the answer.
- Close the gap between intel and action. Make sure that when the research says "feature this," the team can actually produce it in the time the deadline allows.
The firms winning consistently aren't better writers. They started the pursuit earlier — and they could act on what they learned.
Frequently Asked Questions
When is a proposal actually decided?
Often well before submission. Capture-management research finds that 40 to 80 percent of buyers have a preferred firm before proposals are turned in. The written proposal usually confirms a leaning the client already had — which is why pre-RFP positioning matters more than drafting speed.
What is capture planning in AEC?
Capture planning is the work of positioning your firm to win a specific pursuit before the RFP is released. In an AEC context it means understanding the client's funding, priorities, evaluators, and pain points early enough to shape your team, your project selection, and your win themes around them.
What's the difference between client research and a go/no-go decision?
Client research is the input; the go/no-go decision is what you do with it. A go/no-go made without client research is guesswork. Done in the right order, the research tells you whether you are positioned to win — and the go/no-go decision acts on that honestly.
How early should client research start?
Before the RFP — ideally as soon as a pursuit looks real. For known clients with predictable procurement cycles, that can be six to twelve months out. The point is to learn enough to shape your positioning while you still can.
Can a small firm do this without a capture manager?
Yes. Small firms do not need a capture department — they need a repeatable habit: a one-page intel sheet, disciplined debriefs, and an earlier go/no-go conversation. The structure matters more than the headcount.