thinkhard.SI

Prompt repository Sales

Prepare for a discovery call from the prospect's own website

You have a first call with a prospect coming up and want a one-page brief with hypotheses and questions, built from their site and your notes rather than guesses.

Registry no.
SAL-004
Collection
Sales
Laid out for
GPT
Score
10 out of 10
Blanks to fill
8
Added

SalesGPT

The prompt

Score 10 of 10, Excellent

Use it when You have a first call with a prospect coming up and want a one-page brief with hypotheses and questions, built from their site and your notes rather than guesses.

  • {{company}}
  • {{prospect_contact}}
  • {{prospect_company}}
  • {{call_length}}
  • {{our_offer}}
  • {{ideal_customer}}
  • {{prospect_pages}}
  • {{call_notes}}
Full prompt and why it works
Prompt
# Role
You help salespeople at {{company}} prepare for first discovery calls.

# Goal
Write a one-page call brief for my call with {{prospect_contact}} at {{prospect_company}}, so I walk in knowing what they do, what they might need from us, and what to ask. The call is {{call_length}} long.

# Success criteria
- Everything about the prospect comes from the pasted pages or my notes, with the source in brackets, for example [About page] or [my notes].
- Hypotheses about their needs are labeled as hypotheses, and each one points to something specific in the sources.
- The questions are open, specific to this company, and not ones a generic question list would contain.

# Constraints
- Do not use what you remember about this company; it may be out of date or about a different company with a similar name.
- Do not invent revenue, headcount, funding, tools they use or names of people. If it is not in the sources, it goes under Unknowns.
- Keep it under 450 words so it fits on one screen before the call.

# Output
Markdown with these headings:
- What they do: three sentences.
- Why they might need us: up to three hypotheses, each with its evidence.
- Questions to ask: 6 to 8, ordered for a natural conversation, starting broad.
- What to listen for: signs that they are or are not a fit, based on our ideal customer notes.
- Unknowns: what I should find out or confirm.

# Stop rules
If the pasted pages are too thin to support any hypothesis, say so in one line at the top and give the questions anyway.

# Context
<our_offer>
{{our_offer}}
</our_offer>

<ideal_customer>
{{ideal_customer}}
</ideal_customer>

<prospect_pages>
{{prospect_pages}}
</prospect_pages>

<my_notes>
{{call_notes}}
</my_notes>

Variables

{{company}}
Your company name.
{{prospect_contact}}
Who you are meeting and their role.
{{prospect_company}}
The prospect's company.
{{call_length}}
For example 30 minutes.
{{our_offer}}
What you sell and the problems it solves.
{{ideal_customer}}
What a good-fit customer looks like, and the signs of a poor fit.
{{prospect_pages}}
Text copied from their website, such as the home, about, careers or product pages, each labeled with the page name.
{{call_notes}}
How the meeting came about, what they asked for, anything from earlier emails. "None" is fine.

Why it works

  • Background facts. It works from pages you paste and labels each fact with its source, so the brief reflects the company as it describes itself today.
  • What to avoid. Memory about the company is ruled out, along with invented headcount and tools, the details most likely to be wrong and most awkward to repeat on a call.
  • Layout. Markdown sections in OpenAI's suggested order, with the pasted pages in tags at the end, where GPT expects request-specific material.

How to use it

  1. Press Fill in variables and type your details into each blank. What you type stays in this browser.
  2. Copy the filled prompt into the AI tool you already use, such as ChatGPT, Claude or Gemini, or open it in the Grader or the Prompt Tester here.
  3. Before you rely on it, try it on three or four real examples. The Prompt Tester runs them all at once and shows which ones go wrong.
Use it from code

Get this prompt as JSON from the free REST API, with your own text in the blanks, or let an AI assistant fetch it through the MCP server. No key.

curl -s https://thinkhard.si/api/v1/prompts/discovery-call-prep-brief

With the MCP server added to Claude, ChatGPT, Cursor or VS Code, call get_prompt with {"prompt": "discovery-call-prep-brief"}, or just ask for it by name.