Generate Tailored Discovery Questions

When to use it

Use this before a discovery or qualification call when you want questions tailored to this specific account and persona, sequenced so you actually uncover impact and decision process, not just surface-level needs. It earns its keep most for newer reps, complex multi-stakeholder deals, and any call where you tend to walk out with interest but no quantified pain, no timeline, and no map of who decides. Feed it the brief from the account-research prompt for sharper questions.

Do NOT walk into the call and read all 18 questions like a survey, that's an interrogation, and it kills rapport and surfaces nothing real. Pick the 6-8 that fit the natural flow, keep the must-answers non-negotiable, and let the conversation breathe. Also do not use leading questions that beg for a yes; they feel good in the moment and teach you nothing about whether the deal is real.

The principle: discovery is about uncovering quantified impact and the decision process, not confirming that your product sounds nice. The most common failure is leaving with 'pain' but no dollar figure and no idea who signs, which is how deals stall in 'we love it but the timing isn't right.' Good discovery questions are open-ended, tied to this account's actual situation, and push relentlessly toward 'what is this worth' and 'who decides.'

The prompt

Prompt, paste into Claude, ChatGPT, Gemini

You are a sales coach who trains reps to run discovery that uncovers real pain, quantified impact, and the actual decision process, not feature checklists. Your questions are open-ended, conversational, non-leading, and make the buyer think. You tailor every question to the specific account and role; you never produce a generic list.

CONTEXT
What we sell (product + outcome): {{WHAT_WE_SELL}}
Prospect company + what they do: {{PROSPECT}}
Who I'm talking to (role): {{PERSONA}}
Hypothesized pain / why they'd care: {{HYPOTHESIZED_PAIN}}
What I must learn to qualify: {{QUALIFICATION_GOALS}}
Methodology to lean on (MEDDIC, SPIN, Challenger, or none): {{METHODOLOGY}}
What I already know going in: {{KNOWN_CONTEXT}}

TASK
Generate tailored discovery questions, grouped and sequenced for a real conversation.

METHOD
1. Ground every question in THIS account and role, a question that could be asked of any company has failed.
2. Sequence from low-threat context-setting toward higher-stakes impact and process questions.
3. In the impact bucket, push hard toward quantification (dollars, hours, SLA penalties, churn).
4. In the decision bucket, map the committee and the trigger that forces a decision, not just 'who's involved.'
5. For each bucket, name the ONE question I cannot leave the call without answering.

OUTPUT FORMAT (3-4 questions per bucket)
1. CURRENT STATE & PAIN
2. IMPACT & COST OF INACTION (push to quantify)
3. DESIRED FUTURE STATE / SUCCESS CRITERIA
4. DECISION PROCESS, STAKEHOLDERS & BUDGET
5. COMPETITION & ALTERNATIVES (including 'do nothing')
For each bucket, end with: MUST-ANSWER: [the one thing].
Then: FOLLOW-UP PROBES (2): what to ask when an answer is vague or surface-level.

CONSTRAINTS
- Open-ended and conversational. No yes/no leading questions. No jargon-dumping.
- Tailored to the account; cut anything generic.
- If {{METHODOLOGY}} is given, align the impact and decision buckets to it.
- If {{HYPOTHESIZED_PAIN}} or {{QUALIFICATION_GOALS}} is thin, tell me which and ask one clarifying question before generating.

Run it from the terminal

zsh

$# reuse the account brief as input so questions build on research

$cat out/account-brief.txt >> prompts/discovery.md

$# build deal-specific discovery questions (MEDDIC) with the llm CLI

$llm -m claude-sonnet-4-6 < prompts/discovery.md

1. CURRENT STATE & PAIN - Walk me through how routes get planned today, from an order landing to a driver's stops. - With the grocery contract coming online, where does that process start to strain? MUST-ANSWER: Who owns route planning, and how manual is it today? 2. IMPACT & COST OF INACTION - If fuel and driver hours hold through the grocery ramp, what does that do to margin? - What would a 1% improvement in fuel or driver hours be worth annually? MUST-ANSWER: The dollar value of the pain (or a 1% improvement). 4. DECISION PROCESS, STAKEHOLDERS & BUDGET (MEDDIC) - Besides you, who weighs in on a routing change, and who signs? - What has to be true for a decision before the grocery launch? MUST-ANSWER: Economic buyer + the trigger that forces a timeline. 5. COMPETITION & ALTERNATIVES - Have you looked at other options, including building it internally or doing nothing? MUST-ANSWER: The real alternative you're competing against (often 'do nothing'). FOLLOW-UP PROBES - 'Can you give me a recent, specific example of that?' - 'What makes that hard / why hasn't it been fixed already?'

$# pick the 6-8 that fit the flow; keep MUST-ANSWERs non-negotiable

Variables to swap

Variable	What to put there
`{{WHAT_WE_SELL}}`	Your product/service and the outcome
`{{PROSPECT}}`	The prospect company and what they do
`{{PERSONA}}`	Role of the person you're meeting
`{{HYPOTHESIZED_PAIN}}`	Your best guess at why they'd care
`{{QUALIFICATION_GOALS}}`	What you must learn to qualify the deal
`{{METHODOLOGY}}`	A sales methodology to lean on, or 'none'
`{{KNOWN_CONTEXT}}`	What you already know going in, so the model doesn't ask what you've answered

Worked example

Input

Filled-in example

WHAT_WE_SELL: AI route optimization that cuts fuel and driver hours
PROSPECT: Northwind Logistics, regional last-mile delivery
PERSONA: VP Operations
HYPOTHESIZED_PAIN: Manual dispatch can't scale with new contracts; margins under pressure
QUALIFICATION_GOALS: Confirm pain size, budget owner, timeline, current tooling
METHODOLOGY: MEDDIC
KNOWN_CONTEXT: They just signed a new grocery contract and are hiring dispatchers

Output

What the model returns

1. CURRENT STATE & PAIN
- Walk me through how routes get planned today, from an order landing to a driver getting their stops.
- With the grocery contract coming online, where does that process start to strain?
- How much of route planning is manual versus system-driven right now?
MUST-ANSWER: Who owns route planning, and how manual is it today?

2. IMPACT & COST OF INACTION
- If fuel and driver hours hold where they are through the grocery ramp, what does that do to margin on that contract?
- How do you measure a routing miss today, overtime, fuel, SLA penalties?
- What would a 1% improvement in fuel or driver hours be worth annually?
MUST-ANSWER: The dollar value of the pain (or a 1% improvement).

3. DESIRED FUTURE STATE
- A year out, what does 'routing is handled' look like to you?
- What metric would your CEO judge this on?
MUST-ANSWER: The success metric leadership cares about.

4. DECISION PROCESS, STAKEHOLDERS & BUDGET (MEDDIC)
- Besides you, who weighs in on a change to how routing works, and who signs?
- Is there budget set aside for ops tooling, or would this need a case built?
- What would have to be true for a decision to happen before the grocery launch?
MUST-ANSWER: Economic buyer + the trigger that forces a timeline.

5. COMPETITION & ALTERNATIVES
- Have you looked at other options, including building it internally or doing nothing?
- What's kept you on the current process so far?
MUST-ANSWER: The real alternative you're competing against (often 'do nothing').

FOLLOW-UP PROBES
- 'Can you give me a recent, specific example of that?'
- 'What makes that hard / why hasn't it been fixed already?'

Tips

Tips to get more out of it

Don't read all 18, pick 6-8 that fit the natural flow of the conversation and keep the MUST-ANSWERs non-negotiable. A call where you got every must-answer beats one where you asked every question.
Specify your methodology so the impact and decision buckets match how you actually qualify; a MEDDIC team and a Challenger team probe decision process very differently.
Use the follow-up probes live, the second 'why' or 'can you give me an example' is almost always where the real, quantified pain surfaces. Surface answers are where weak reps stop.
Fill KNOWN_CONTEXT honestly so the model doesn't waste questions asking what you already know, and so the questions build on, rather than repeat, your research.
After the call, paste your notes into the follow-up prompt to turn answers into next steps and a mutual action plan fast, while it's fresh.
If a deal stalls weeks later, re-run this with what you now know, it usually exposes the must-answer you never actually got (often the economic buyer or the real timeline trigger).
Push the model if questions feel generic: 'these could be asked of any logistics company, make them specific to a regional carrier ramping a grocery contract.'

Get new prompts in your inbox

The best GTM prompts and AI use cases, every couple of weeks.

Subscribe →