Deciding whether an inbound enquiry is worth a meeting
Qualifying an inbound enquiry in two minutes
A prompt that researches a company against your own criteria and tells you why it scored what it scored.
The video for this one is not recorded yet.
The prompt below is finished and works on its own. Set videoUrl, published and duration in the entry to put the recording here.
The prompt
Deciding whether an inbound enquiry is worth a meeting
You are qualifying an inbound enquiry. I will give you a company name and website, and my
qualification criteria. Research the company and tell me whether it is worth a meeting.
My criteria:
[Replace this block with your own. Be specific. For example: business services, ten to two
hundred people, United Kingdom or Ireland, has raised outside capital or is owner managed
and profitable, and has a repetitive internal process that a small team currently does by
hand.]
Return exactly this structure.
1. What the company does, in two sentences, written so that somebody who has never heard
of it would understand the business. Not their marketing language.
2. Size and stage, with the evidence you used. If you are estimating, say that you are
estimating and say from what.
3. Score against each of my criteria individually. For each one: met, not met, or unclear,
followed by the specific evidence. Never mark a criterion as met without evidence.
4. An overall recommendation of meet, decline, or ask first.
5. If the recommendation is meet: the single question I should ask at the start of the
call that would most change my view of the opportunity.
6. If the recommendation is ask first: the exact question to send back by email before
committing to a call.
7. The two things you could not establish that would most change the recommendation.
Constraints:
- Use only what you can find and cite. If you cannot verify something, put it in point
seven rather than guessing at it.
- Do not soften a decline. If the company does not fit, say so plainly and say which
criterion it failed.
- Do not tell me what the company might do in future. Assess what it is now.Qualification is a filter, and filters only work if they reject things. The common failure in an automated qualification step is a model that finds something positive to say about every company, because that is what it has been trained to do.
Two parts of this prompt exist to counteract that. Criteria are scored one at a time with evidence attached, rather than assessed as an overall impression. And the model is required to state what it could not establish, which gives the unclear cases somewhere to go other than into a generous score.
How to use it
The criteria block is the part that matters. Write it in your own words, as specifically as you can, and rewrite it after every ten enquiries you have judged for yourself. Your criteria will be wrong at first. Seeing them scored against real companies is how they improve.
What to watch for
If the model returns meet for almost everything, the criteria are too loose rather than the model being too generous. Tighten them and run the same ten companies again.
If a prompt gets you most of the way there and you want the rest of it built properly, that is the work.