A discovery call should end with a decision.

Prepare a hypothesis, not a pitch deck

Before the call, read the brief and inspect any materials the prospect has shared. Write down the likely problem, the evidence you have, and the three unknowns that could change the project. Do not build a free custom solution just to prove you can contribute. Bring one relevant example and explain what it does and does not demonstrate.

Minutes 0–5: understand why now

Ask: What prompted this conversation? What is happening today? Who experiences the problem? What would a useful improvement let them do? Listen for observable events rather than broad wishes such as “better marketing.” If a client wants an AI chatbot, the underlying need might be faster answers from a reliable knowledge base.

Minutes 5–12: inspect the work

Ask what exists already, who owns it, and whether it is ready to use. Find out how approval works and whether there are accessibility, brand, data, or technical requirements. When AI might help, ask which tools and data uses are permitted. You can discuss the process without collecting sensitive material during an exploratory call.

For a newsletter pilot, clarify whether the interview is recorded, who approves claims, how quickly feedback arrives, and whether publication is part of your scope. A deadline that ignores those dependencies is not a reliable promise.

Minutes 12–17: discuss constraints

Ask about the decision process, intended timing, and the budget available for this piece of work. If the budget is below your workable scope, reduce the deliverable or decline. Do not silently remove review or accept a larger promise because you hope AI will make it effortless.

“For that budget, I could focus on one approved newsletter rather than the full monthly content program. Would solving that smaller piece still be useful?”

Minutes 17–20: choose the next step

Summarize the problem, likely deliverable, missing inputs, and who will do what next. The result may be a proposal, a paid discovery phase, a referral, or a decision not to proceed. If key uncertainties remain, name them instead of guessing a fixed price.

Send a short recap through the appropriate channel. Separate agreed facts from items that still need approval. Ask permission before recording or transcribing. Do not paste identifiable call notes into an external AI tool without the necessary permission.

Review your own performance

After the call, ask whether you learned enough to scope the work and whether the prospect understood your next step. If the same question repeatedly remains unanswered, improve the agenda. The goal is a useful decision, not a perfect sales performance.

Copy the discovery recap and turn agreement into a scope draft.