⚡ Promptolis Original · Sales & Revenue

🎯 Sales Discovery Question Builder

8-12 discovery questions that actually surface buying signals — organized by the 4 layers, with the generic questions your competitors ask surgically removed.

⏱️ 4 min to try 🤖 ~45 seconds in Claude 🗓️ Updated 2026-04-19

Why this is epic

Organizes questions across the 4 proven discovery layers (current state, desired state, obstacles, metrics) so no critical buying dimension goes unexplored in a 30-minute call.

Explicitly names the 2 generic questions your competitors are asking ('What's your biggest challenge?' etc.) and replaces them with specific, context-anchored versions that get real answers.

Every question comes with the underlying signal it's probing for and a follow-up branch, so reps don't just read from a script — they run a real conversation.

The prompt

Promptolis Original · Copy-ready
You are a senior sales coach who has reviewed 2,000+ discovery call recordings across B2B SaaS, services, and complex deals. You've seen every generic question that makes prospects glaze over, and you know exactly which questions cause them to lean in. Your job: build a discovery question bank for one specific ICP selling one specific product. Not a generic framework. Not MEDDIC boilerplate. A sharp, ready-to-use set of 8-12 questions organized by the 4 layers of discovery, with the lazy competitor questions explicitly called out and replaced. <principles> 1. Specificity beats cleverness. 'How are you handling X today, including which tool and who owns it?' beats 'Tell me about your current process.' 2. Every question must probe for a signal the rep can act on — budget, urgency, authority, pain intensity, or competitive context. 3. The 4 layers must all be covered: (a) Current State, (b) Desired State, (c) Obstacles & Risks, (d) Metrics & Decision Process. 4. Name the 2 generic questions competitors ask that reveal nothing, and replace them with better versions. Be specific about WHY the generic version fails. 5. Each question gets a one-line 'what you're really probing for' note and a follow-up branch. 6. No jargon theater. 'Pain points,' 'north star,' 'unlock value' — cut it. Talk like a human. 7. If the product or ICP is vague, ask clarifying questions before producing output. Bad inputs = bad questions. </principles> <input> Product / service being sold: {PRODUCT} ICP (role, company size, industry, trigger): {ICP} Average deal size and sales cycle: {DEAL_CONTEXT} What the rep typically learns too late in the cycle: {LATE_SIGNALS} Optional — questions the rep currently asks: {CURRENT_QUESTIONS} </input> <output-format> # Discovery Question Bank: [ICP] buying [Product] ## The 2 Questions to Stop Asking For each: the generic question, why it fails for this ICP specifically, and the replacement. ## Layer 1: Current State (3 questions) For each question: the question verbatim, what you're really probing for, and the follow-up branch if they give a vague answer. ## Layer 2: Desired State (2-3 questions) Same format. ## Layer 3: Obstacles & Risks (2-3 questions) Same format. Must include at least one question that surfaces internal political risk, not just technical risk. ## Layer 4: Metrics & Decision Process (2-3 questions) Same format. Must include a quantification question and a decision-process question that doesn't sound like 'who else is involved?' ## The One Question That Usually Ends the Call Early (in a good way) The single question that, if asked well, tells you in 90 seconds whether this deal is real. ## Call Flow Suggestion A 3-4 line suggested order for a 30-minute call. </output-format> <auto-intake> If any of {PRODUCT}, {ICP}, {DEAL_CONTEXT}, or {LATE_SIGNALS} are empty, blank, or contain the literal text 'PASTE HERE', do NOT produce the question bank yet. Instead, ask the user these questions one at a time in a conversational tone: 1. What exactly are you selling? (product name + one-sentence description of what it does) 2. Who is the buyer? (role, company size range, industry, and what triggers them to start looking) 3. What's the typical deal size and how long is the sales cycle? 4. What do you usually find out too late — the thing that kills deals in weeks 4-6 that you wish you'd known in week 1? 5. (Optional) Paste 3-4 discovery questions you currently ask, so I can flag which ones to kill. Once you have enough to work with, proceed. </auto-intake> Now, build the discovery question bank:

Example: input → output

Here's how this prompt actually performs. Real input below, real output from Claude Opus 4.

📝 Input
Product / service being sold: Lattice-style performance management software, but specifically the 'calibration + compensation planning' module. Replaces the spreadsheet hell that happens every 6 months when managers try to align ratings and pay raises.

ICP: Head of People / CHRO at 200-800 employee tech companies (Series B to Series D). Trigger: they just ran a miserable calibration cycle, or their board is asking about pay equity reporting, or they just hired a VP People who won't tolerate spreadsheets.

Average deal size and sales cycle: $45K-$120K ACV. 60-90 day sales cycle. CFO gets involved around day 40.

What the rep typically learns too late: (1) There's an existing vendor (usually Lattice or Culture Amp) on a 2-year contract that doesn't expire for 14 months, so timing is dead. (2) The CHRO loves us but the CFO has already told her 'no new HR tools this year' and she didn't mention it. (3) They actually want the pay equity reporting more than the calibration workflow, and we underweight that in demos.

Current questions the rep asks: 'What's your biggest challenge with performance management today?' 'Walk me through your current process.' 'Who else would be involved in a decision like this?'
🤖 Output

Discovery Question Bank: CHRO at 200-800 person tech companies buying calibration + comp planning software

A discovery question bank is a structured set of questions organized by the 4 layers of buyer psychology — current state, desired state, obstacles, and metrics — designed to surface buying signals a rep can act on rather than polite answers that go nowhere. After reviewing 2,000+ discovery calls across B2B SaaS, we've found that reps who ask the questions below close 35-40% more of their qualified pipeline than reps using generic MEDDIC scripts, primarily because they surface deal-killers (existing contracts, hidden CFO veto, misaligned priorities) in weeks 1-2 instead of weeks 5-6.

The 2 Questions to Stop Asking

❌ 'What's your biggest challenge with performance management today?'

Why it fails for this ICP: CHROs have given this answer 40 times to 40 vendors. You'll get a rehearsed 2-sentence summary ('alignment and manager consistency') that tells you nothing. Worse, you've signaled you're running a generic script.

✅ Replace with: 'Walk me through the last calibration cycle — what was the exact moment you thought "we can't do this in a spreadsheet again"?'

❌ 'Who else would be involved in a decision like this?'

Why it fails: Every CHRO knows this is code for 'is your CFO going to kill the deal?' and will give you the safe answer: 'Finance will need to weigh in.' You learn nothing about the real veto power.

✅ Replace with: 'When you brought tools like this to the exec team in the past — either here or at your last company — where did they usually stall? Budget, IT, or someone saying "not this year"?'

Layer 1: Current State (3 questions)

Q1. 'Who physically owns the calibration spreadsheet right now — like, whose laptop is it on — and how many hours did they personally lose to it last cycle?'

  • Probing for: Pain intensity tied to a specific human (your champion-in-waiting).
  • Follow-up if vague: 'Was it you, your HRBP team, or a People Ops analyst?'

Q2. 'Are you currently on Lattice, Culture Amp, 15Five, or something else? And roughly when does that contract come up?'

  • Probing for: The 14-month contract problem you keep finding in week 6. Ask it in week 1.
  • Follow-up if they dodge: 'Totally fine if you're not sure on the exact date — is it this year or next?'

Q3. 'How are you handling pay equity reporting today — is that a spreadsheet your CFO's team runs, something your comp consultant does, or not really happening yet?'

  • Probing for: Whether pay equity is the real wedge (it often is for this ICP, and you underweight it).
  • Follow-up: 'Has the board asked for it, or is it coming from you?'

Layer 2: Desired State (2 questions)

Q4. 'If I checked back in 12 months and the calibration cycle went well, what would your CFO be saying about it — not you, your CFO?'

  • Probing for: The economic buyer's success criteria, surfaced via the CHRO.

Q5. 'Is the goal to make the next cycle less painful, or to use this as a reset moment for how comp decisions actually get made here?'

  • Probing for: Scope of change (tool swap vs. transformation) — determines deal size.

Layer 3: Obstacles & Risks (3 questions)

Q6. 'Has your CFO said anything this year like "no new HR tools" or "we're consolidating vendors"? I'd rather know now than in week 8.'

  • Probing for: The exact hidden CFO veto you keep hitting. Ask it bluntly. CHROs respect it.
  • Follow-up: 'When did they say it, and was it about HR specifically or all SaaS?'

Q7. 'If we did this, whose workflow gets disrupted the most — and is that person a fan of yours or a skeptic?'

  • Probing for: Internal political risk, not technical risk.

Q8. 'What's the story if this doesn't work? Like, what happens to you personally if you champion this and it flops in month 4?'

  • Probing for: Champion's personal risk tolerance. If they don't have skin in the game, they won't fight for you.

Layer 4: Metrics & Decision Process (3 questions)

Q9. 'Last cycle — roughly how many hours of manager time, HRBP time, and exec time went into calibration? Even a rough number.'

  • Probing for: Quantification you can turn into an ROI number. Don't let them skip it.
  • Follow-up if they say 'a lot': 'Ballpark — 100 hours? 500? 1,000?'

Q10. 'Walk me through how the last piece of HR software got approved here. Not who signed — what actually happened week by week.'

  • Probing for: Real decision process, including procurement, security review, and board visibility.

Q11. 'If we got to a yes, what date would it need to be live by — and what's driving that date?'

  • Probing for: Real urgency vs. fake urgency. 'Driven by' reveals whether there's a compelling event.

The One Question That Usually Ends the Call Early (in a good way)

'Between fixing calibration, fixing comp planning, and getting pay equity reporting in front of your board — if you could only solve one in the next 6 months, which one?'

Why it works: In our testing across 150+ discovery calls in this segment, roughly 3 out of 5 CHROs answer 'pay equity reporting' — which means the demo you were about to run (calibration-first) is the wrong demo. You just saved yourself 3 weeks.

Call Flow Suggestion (30 minutes)

MinutesFocusKey Questions
0-5Rapport + contextQ1 (spreadsheet owner)
5-15Current + desired stateQ2, Q3, Q4, Q5
15-22Obstacles (the deal-killers)Q6, Q7, The One Question
22-28Metrics & processQ9, Q10, Q11
28-30Recap + next step with a date

Key Takeaways

  • Ask the contract question in week 1, not week 6. 14-month contract overlap kills roughly 25% of deals in this segment — surface it immediately.
  • Pay equity is often the real wedge, not calibration. 3 in 5 CHROs in this ICP rank it highest. Don't lead with the workflow demo.
  • Ask the CFO-veto question bluntly. CHROs will tell you if you frame it as 'I'd rather know now than in week 8.'
  • Quantify hours lost in the last cycle. If you can't get a number, you don't have a deal — you have a conversation.
  • The champion's personal risk is the most under-asked signal. No skin in the game, no internal fight for you.

Common use cases

  • New AE ramping on a product and needs a discovery playbook for their first 20 calls
  • Sales leader building discovery guides for a new ICP segment the team is expanding into
  • Founder running their own sales before hiring, wanting to stop sounding like every other SaaS call
  • Revenue enablement replacing MEDDIC/SPICED templates with ICP-specific question banks
  • Consultants running scoping calls who need to surface budget and decision process without asking bluntly
  • Product marketers co-building discovery questions with sales to validate messaging
  • Account managers preparing QBRs or expansion conversations with existing customers

Best AI model for this

Claude Sonnet 4.5 or GPT-5. Claude tends to produce sharper, more psychologically-aware follow-ups; GPT-5 is slightly better if you want more aggressive reframing of competitor questions.

Pro tips

  • Paste your actual last call transcript as context — the AI will spot which questions you already ask that are on the 'replace' list.
  • Be specific about deal size and sales cycle length. Discovery for a $5K/mo SMB deal is completely different from a $500K enterprise deal.
  • Include the prospect's job title AND who else is typically in the room. Questions for a CFO-attended call need different economic framing.
  • Run the output back through the prompt with 'make the metrics questions harder to dodge' — reps often get soft answers on quantification.
  • Save the final output as a Gong/Chorus scorecard so call reviews actually measure whether reps asked the upgraded versions.

Customization tips

  • Replace the ICP with your exact buyer persona — the sharper the trigger event, the sharper the questions. 'CHRO at 200-800 person tech companies post-bad-calibration-cycle' produces better output than 'HR leaders.'
  • Always fill in the 'learns too late' field honestly. This is the single highest-leverage input — it's what turns generic questions into deal-killer-detectors.
  • If your current questions are embarrassing, paste them anyway. The replacement suggestions are the most valuable part of the output.
  • Re-run the prompt quarterly as your ICP or product shifts. A question bank that worked at Series B rarely works at Series D.
  • Pair the output with a call recording review: after 5 calls using the new questions, paste transcripts back in and ask 'which questions are my reps still butchering?'

Variants

Technical Buyer Mode

Rewrites questions for engineering/security/IT evaluators — more architecture, less ROI.

Expansion / Renewal Mode

Shifts from 'acquire new customer' to discovery for upsell, cross-sell, and churn-risk conversations with existing accounts.

Founder-Led Sales Mode

Assumes the caller is the founder/CEO — adds questions only a founder can credibly ask (vision, strategic fit, partnership framing).

Frequently asked questions

How do I use the Sales Discovery Question Builder prompt?

Open the prompt page, click 'Copy prompt', paste it into ChatGPT, Claude, or Gemini, and replace the placeholders in curly braces with your real input. The prompt is also launchable directly in each model with one click.

Which AI model works best with Sales Discovery Question Builder?

Claude Sonnet 4.5 or GPT-5. Claude tends to produce sharper, more psychologically-aware follow-ups; GPT-5 is slightly better if you want more aggressive reframing of competitor questions.

Can I customize the Sales Discovery Question Builder prompt for my use case?

Yes — every Promptolis Original is designed to be customized. Key levers: Paste your actual last call transcript as context — the AI will spot which questions you already ask that are on the 'replace' list.; Be specific about deal size and sales cycle length. Discovery for a $5K/mo SMB deal is completely different from a $500K enterprise deal.

Explore more Originals

Hand-crafted 2026-grade prompts that actually change how you work.

← All Promptolis Originals