ChatGPT prompt for qualifying an inbound lead

This is the prompt we run when a new inbound lead lands and we have to decide whether it deserves a call. Its core rule: it infers no missing fact. A criterion is only a “match” if it is written in the lead’s message; everything else goes to UNKNOWN.

The prompt to copy

New inbound lead. What we sell: {one line}. Our ideal client criteria: {list}. Lead info: {paste}.
Produce:
1. Fit table — each criterion: match / no match / UNKNOWN, citing only what's in the lead info. Do not infer facts that aren't there.
2. Red flags, if any.
3. The 3 questions that would resolve the biggest unknowns.
4. Recommended next step — book a call / reply with questions / politely decline — with a draft of that reply, under 100 words.

Replace the fields in curly braces (what you sell, your criteria, the lead info), then paste the prospect’s message exactly as it is.

When to use it

The moment an inbound lead hits the inbox, before you block a slot. The temptation is to read the message optimistically and fill in the missing cells in your head. This prompt does the opposite: it separates what the lead said from what you are hoping, and you come out with the three questions to ask before you sink an hour into a call.

What you get back

1. A fit table criterion by criterion, backed only by the facts in the message.
2. The red flags, if any.
3. The 3 questions that would resolve the biggest unknowns.
4. A recommended next step, with the reply already drafted.

Real example: inbound message → fit table

Input: a lead received by the Brightside agency, an 8-person shop. What it sells: “brand identity refresh for food-and-drink SMEs”. Ideal-client criteria: budget above 8,000, a named decision-maker, a realistic timeline. The lead’s message, as-is:

Hi, we love your work for Fernhill Coffee. We make jams and our packaging looks dated. We'd like to talk about a refresh. Are you free this week?

Output, the part that matters:

Criterion	Verdict	What backs it
Food-and-drink sector	Match	“we make jams”
Budget above 8,000	UNKNOWN	no figure mentioned
Named decision-maker	UNKNOWN	signer not stated
Realistic timeline	UNKNOWN	“this week” is about the call, not the project

Notice what the AI did not do: it did not assume a budget because the brand “loves your work”, nor promote the message’s author to decision-maker. Three cells out of four stay UNKNOWN, which becomes exactly the list of questions to ask. That is the prompt working, not failing. The human checkpoint, deciding to call or to decline, stays in your hands.

The full worked example (complete message, fit table, drafted reply) ships inside the kit, with the other prompts in the sales phase.