A discovery call that jumps straight to pitching your product almost always loses to one that spends most of its time asking good questions first. AI is well suited to drafting those questions, but the trap is asking for “discovery call questions” and getting a generic list that could apply to any product. The fix is prompting for questions built around your specific offer and a proven questioning sequence.
What Question Sequence Should You Ask AI to Follow?
SPIN selling, a methodology developed by Neil Rackham, structures discovery around four question types asked in order: Situation questions establish context, Problem questions surface pain, Implication questions build urgency around that pain, and Need-payoff questions get the prospect to articulate the value of solving it themselves. Naming this sequence explicitly in your prompt produces far better output than asking for “good discovery questions” in general:
“Write discovery call questions for [your product/service] following the SPIN framework: 2-3 Situation questions to understand their current setup, 2-3 Problem questions to surface pain points, 2-3 Implication questions that explore the cost of not solving the problem, and 2-3 Need-payoff questions that get the prospect to describe the value of a solution in their own words.”
How Do You Prompt for Questions Specific to Your Buyer, Not Generic Ones?
Generic discovery questions (“What are your biggest challenges?”) get generic, guarded answers. Give the AI real context about who you’re calling and what you sell before asking for questions:
“I’m selling [specific product/service] to [specific buyer role, e.g. ‘IT directors at mid-size manufacturing companies’]. The problem we solve is [specific problem]. Write Problem and Implication questions that reference the actual workflow this buyer deals with day to day, not generic business pain points.”
What’s a Good Prompt for Following Up on a Vague Answer?
The real skill in discovery isn’t the first question, it’s the follow-up when a prospect gives a surface-level answer. Prepare for this before the call:
“A prospect just answered my discovery question with: [paste a vague, typical answer like ‘we’re fine, just looking at options’]. Write three follow-up questions that dig past that surface answer to find the real motivation or pain behind it, without sounding pushy.”
How Should You Prompt AI to Prep for the Call Itself?
Beyond the questions themselves, use AI to prepare a one-page call brief so you’re not reading from a script live:
“Turn this list of SPIN discovery questions into a one-page call outline I can glance at during the call: group the questions under Situation, Problem, Implication, and Need-payoff headers, and add a one-line reminder under each section about what I’m listening for in the prospect’s answer.”
If the call surfaces an objection partway through, have your follow-up prompts ready — see AI Prompts for Sales Call Scripts That Handle Common Objections for that part of the conversation.
Frequently Asked Questions
How many discovery questions should a single call actually cover?
Fewer than you’d think — six to ten well-chosen questions, with room for follow-ups, usually fills a 30-minute call. Ask AI for a focused set tied to your SPIN sequence rather than a long master list, and pick the ones most relevant to that specific prospect beforehand.
Should every discovery call use the exact same questions?
No. Use AI to generate a reusable question bank organized by SPIN category, then pick and lightly adapt the ones that fit each specific prospect’s role and situation, rather than reading an identical script on every call.
What comes after a strong discovery call?
The pain and value the prospect articulated during Implication and Need-payoff questions should show up directly in your next step — typically a proposal. See AI Prompts for Writing Sales Proposals That Get Signed Faster for turning discovery insights into a proposal that references what the prospect actually told you.
Further reading: Gong’s breakdown of the SPIN selling framework covers the full question-type methodology in more depth.
