Best ChatGPT Prompt for Discovery Call Questions
Pain, authority, budget, timeline — in a conversation that doesn't sound like a script.
The Prompt
You are a senior B2B sales leader who has run thousands of discovery calls and trained AEs to do the same. Generate discovery questions for an upcoming call. What we sell: [ONE SENTENCE] Who the call is with (role + company + size): [SPECIFIC] Where they came from (inbound / outbound / referral / event): [SPECIFIC] What I already know about their situation: [WHATEVER YOU HAVE] Average sales cycle and deal size: [E.G. “45 days, $25K ACV”] What disqualifies a deal for us: [BE SPECIFIC — company size, tech stack, geography, etc.] Requirements: - 12-15 questions organized into: Context, Pain, Impact, Decision Process, Budget, Timeline - Open-ended (no yes/no questions), specific (not “tell me about your challenges”) - For EACH question, give me: the question itself, what a strong answer sounds like, what a red-flag answer sounds like, and the follow-up to ask if the answer is weak - Include 2-3 questions designed to uncover competing solutions they're evaluating - Include 1 question that will surface internal political risk (champions, blockers, executive sponsor) - Do NOT include “what keeps you up at night” or “if you could wave a magic wand” or any other sales cliché
How to Use This Prompt
- Specify their role — you ask a VP very different questions than an IC champion
- Be honest about your disqualifiers — the questions should screen them in or out, not just qualify in
- Note where they came from — inbound vs outbound calls run differently in the first 5 minutes
- Paste into ChatGPT or Claude
- Pick 8-10 from the list for the actual call; don't try to ask all 15
Discovery calls fail when reps ask too many questions and never listen. A good call asks 8-12 questions and spends 60% of the time letting the prospect talk. The list is for prep; don't read from it.
Example Output
Task: Discovery questions for a 30-min call with a VP of Engineering at a 200-person SaaS company about an observability product
Response:
CONTEXT (5 min):
1. What does your current observability stack look like?
Strong answer: names 3-5 specific tools, knows costs, owns budget.
Red flag: “we have a few things, the team handles it.”
Weak-answer follow-up: “Who on your team would know off the top of their head?”
2. Walk me through the last on-call incident your team handled. What happened?
Strong answer: specific incident, named root cause, named time-to-resolution.
Red flag: vague, can't recall a specific one.
Weak-answer follow-up: “What's the most recent one that lost someone sleep?”
PAIN (8 min):
3. When something breaks in production, what's the first place your team looks?
... (continues for all 12-15 questions)
DECISION + POLITICAL:
12. Who else, besides you, would need to weigh in on a decision like this?
Strong answer: names 2-3 people by name and role.
Red flag: “just me” (rarely true at 200-person company — signals low engagement
or hidden blockers).
Tips to Get Better Results
- Ask, then shut up. The pause after a hard question is where you learn things. Sales reps lose deals by filling the silence with the next question.
- Get to budget by call 1 or 2. If you finish the discovery without knowing whether they can buy what you sell, you've wasted both sides' time.
- Surface competition. “Who else are you looking at?” is a fair question by minute 20. Reps who don't ask end up surprised in the demo.
- Screen out, too. Discovery is also disqualification. If the call surfaces a hard disqualifier (no budget, wrong size, wrong tech stack), end it gracefully instead of forcing it forward.