Discovery: questions that uncover the need
AI Sales 13 min
0/19
Business problem
Salespeople turn discovery into a presentation: 'let me tell you about our product'. The client doesn't feel heard, and the salesperson doesn't gather information for the offer. Closed questions ('do you have a problem with X?') give 'yes/no' answers and zero context. The result: an offer shot in the dark.
Proven method
Discovery in 5 blocks:
- OPENING: 'Tell me how [area] looks at your place now' — one topic, open-ended,
- PAIN DISCOVERY: the current process (how do you do it today?), the problem's cost (what happens when [problem]?), the emotion (what bothers you most about it?),
- IMPACT: 'If [problem] disappeared, what would change?' 'What would that give the business?' — that's the foundation of the offer,
- QUALIFICATION: budget (is there one?), authority (who decides?), timeline (when do they want to implement?), current solution (what do they use now?), desired outcome (what's the goal?),
- CLOSE: 'Did I miss anything?' — the only closed question at the end. The 80/20 rule: in discovery you listen 80% of the time. AI:
- BEFORE: preparing questions by context (the brief from module
- ,
- DURING: notes/transcription (if you have consent),
- AFTER: analyzing answers, identifying missing information (what I didn't ask), preparing next questions and next steps.
Market example
- BAD
- Improvised general pitch lacking buyer context or verified data.
- GOOD
- Instead of 'Do you have a problem with inquiry handling?' you ask: 'Tell me how your customer inquiry handling looks now — from the first email to a response?' (opening). 'What happens when inquiries grow in the season?' (the problem's cost). 'If responses took 2h instead of 24h, what would that change in sales?' (impact). 'Who decides on handling tools?' (authority). 'When would you like to implement something?' (timeline). These answers are material for an offer, not a general 'yes'.
Ready-to-use prompt
You are a discovery coach. Context: I sell [PRODUCT]. Client: [COMPANY + CONTEXT FROM THE BRIEF]. Conversation area: [AREA — e.g. customer service, pipeline]. Task: Prepare a discovery plan: 1. An opening question (open-ended, 1 topic) 2. Pain discovery: 3 questions (current process / the problem's cost / emotion) 3. Impact: 2 questions (what would happen if the problem disappeared — for them and the business) 4. Qualification: questions about budget / authority / timeline / current solution / desired outcome 5. A closing question 6. 'What to look for in answers' — what to watch for (interest signals, objections, missing info) 7. After the conversation: how AI will help analyze the answers and what to do with missing information Format: 7 sections.
Model rekomendowany: Claude 3.5 Sonnet / GPT-4o Copy and paste into ChatGPT, Claude or Perplexity
ALTERNATYWNE SCENARIUSZE
Variations & alternative scenarios
Short conversation version (30 min): pick the 5 most important questions.
Technical product version: questions about the current stack and migration.
Practical exercise
Write 8 discovery questions for your product (2 opening, 3 pain, 2 impact, 1 qualification). Mark which are open-ended. Prepare 1 question you'll use in the nearest conversation.
Zapisano w przeglądarce