Good sales discovery questions help you understand a buyer's current workflow, the consequence of leaving it unchanged, the alternatives they can use and the next decision they are actually ready to make. A founder-led discovery call should produce an evidence-based next step, including a respectful pause when there is no suitable buying project.
This guide gives you a pre-call brief, an original question bank, a qualification worksheet and a buyer-confirmable recap. It is for an active B2B sales conversation. If you are still investigating whether a problem exists across a market, start with customer discovery interviews; do not turn exploratory research into an undisclosed sales pitch.
Prepare facts and unknowns before writing questions
A useful pre-call brief is short enough to use during the conversation. Record the facts you can support, the source for each fact, and what you need to learn. A contact's job title does not prove purchasing authority. A recent funding announcement does not prove available budget for your product.
GitLab's public discovery playbook connects preparation, qualification and learning about customer value and process. Its qualification questions are a primary example of organizing the conversation around the opportunity. The brief and prompts here are original, adaptable tools rather than a copied script or a universally required sales methodology.
Conversation
Company: [name]
Participant: [name and role]
Why the call was requested: [their words, if available]
Starting context
Current workflow: [fact and source, or unknown]
Current alternative: [fact and source, or unknown]
Potential fit: [explicit hypothesis]
Three things to clarify
1. Current work and its consequences: [question]
2. Evaluation or change process: [question]
3. A useful next decision: [question]
Boundaries
Information we may discuss or record: [permission limits]
Reason to pause: [clear mismatch or missing prerequisite]Prepare a small number of relevant follow-ups rather than displaying a long questionnaire. You can learn more from asking how the last instance of a problem unfolded than from asking a buyer to rate an abstract pain from one to ten.
Questions about the current workflow
Ask how the work happens today before proposing a better system. Learn who starts it, where information moves, where someone makes a decision and what the final output needs to be.
- Could you walk me through the last time your team did this work?
- What triggered it, and what happened first?
- Which people or systems did you rely on along the way?
- Where did someone need to interpret information or decide what happened next?
- What did a successful result look like to the person receiving the work?
Follow with an example request: “What did you do when that happened?” Distinguish a frequent, costly problem from an unusual inconvenience. Do not silently fill in a missing step with your product's preferred workflow.
Questions about consequences and priority
- When this process works poorly, what happens next?
- Who notices the consequence, and how do they respond?
- What have you already tried to improve it?
- How does this compare with the other work your team could prioritize?
- Is there a real event driving the timing, or are you exploring options?
If a buyer mentions saved time, ask how they currently observe or estimate that time. Keep the answer as their estimate unless there is a defined measurement. Do not multiply an uncertain estimate into a promised annual ROI or attribute a business outcome to your product before a test.
Questions about alternatives and evaluation
The current spreadsheet, an internal process and doing nothing are genuine alternatives. Your conversation needs to understand why each is acceptable or inadequate, not merely identify another software vendor.
- What is useful about your current approach that you would want to preserve?
- Which other approaches are you seriously considering?
- What would you need to see to decide that a change is worthwhile?
- What information would make an option unsuitable?
- Who would use the result, and who would need to accept the work of changing?
- How are purchases or experiments like this normally evaluated here?
Ask about budget and approval process when relevant, but allow “not agreed yet” to remain an answer. Someone's willingness to attend a demo is not a commitment to purchase. If security, privacy or implementation questions arise, identify the responsible person and the information needed; do not improvise an official commitment.
Questions that find the next useful decision
- What decision could your team reasonably make after the next step?
- Which example or input would let us test that question?
- Who needs to participate for the result to be meaningful?
- What would count as a useful result, and what would tell us to stop?
- Would a follow-up be helpful now, or should we pause until a prerequisite is resolved?
A narrower next step is often clearer than a full demo. You might test whether an approved sample can answer a specific question. Agree the output and acceptance criteria first. Do not ask the buyer to share confidential records simply to make the demo impressive.
Worked example: a demo request without a buying project
Illustrative example: A founder at fictional Cedar Signals receives a demo request from an operations lead at fictional Dockside Services. The lead wants less manual reporting but says leadership has not assigned an owner or budget to a replacement project. All names, conversation details and outcomes below are invented for teaching.
| Topic | What the buyer reports | What the founder records |
|---|---|---|
| Current workflow | Two teams update a shared spreadsheet before a weekly review | Reported process, not independently observed time savings |
| Consequence | The review sometimes starts with conflicting versions | Ask for an approved example of a conflict and how it was resolved |
| Priority | No agreed replacement project exists | Budget, sponsor and purchase timing remain unknown |
| Next decision | Lead wants to see whether a clearer source record would help | Propose a small approved-sample review, not a fabricated close date |
| Pause condition | No permission to share live customer records | Use a synthetic example or pause; do not request a private export |
A reasonable recap is: “We can review a synthetic example and decide whether a source-record approach is relevant enough to discuss internally.” It is not: “You confirmed an urgent purchase this quarter.” The first statement helps both sides learn; the second makes your pipeline look more advanced than the evidence supports.
Discovery qualification and pause record
Conversation
Opportunity or conversation ID: [identifier]
Date: [actual date]
Participants: [names and roles]
Permission and source notes: [approved reference]
What we know
Current work: [specific process]
Consequence: [reported, observed or estimated, with source]
Current alternative: [detail]
What works about it: [detail]
Change priority: [confirmed, hypothesis or unknown]
Evaluation criteria: [buyer-confirmed or proposed]
Participants' roles: [confirmed roles, not inferred authority]
Budget process: [known process or unknown]
Implementation owner: [accepted, not assigned or unknown]
Next decision
Question to resolve: [specific question]
Smallest useful step: [bounded activity and permitted input]
Acceptance evidence: [what the step should show]
Pause condition: [mismatch, no permission or missing prerequisite]
Open question: [question]
Question owner: [person who can clarify]
Recap confirmation: [source or still unconfirmed]
Pause log
Reason: [specific, respectful explanation]
What would need to change: [condition]
Follow-up preference: [buyer-approved date, no follow-up or unknown]Do not force a weighted score onto unknowns. A visible gap is more useful than a high qualification score produced by assumed budget, authority and urgency. If your organization uses a formal qualification method, map these observations into it without changing the underlying evidence.
Send a recap the buyer can correct
Subject: What we learned and a proposed next step
Hi [name],
Here is my understanding of [initiative]. Today, [current workflow in the buyer's terms]. The issue worth checking is [specific consequence]. We still need to clarify [priority, owner, process or other gap].
The proposed next decision is [one clear question]. To answer it, we could [bounded step] using [permitted input].
We would look for [acceptance evidence] and stop if [pause condition].
Please correct anything I misunderstood. If this is not a priority now, we can pause rather than add another meeting.Practice the conversation with Aurora, Claude or Codex
Practice can help you notice leading questions. It cannot tell you what real customers want. Use a fictional scenario or approved notes, and keep the assistant from inventing facts about the prospect.
Task
Help me prepare for a founder-led B2B discovery call.
Tool: [Aurora AI in Finta, Claude or Codex]
Inputs
Permitted facts: [paste approved context]
Source references: [paste approved references]
Discussion and recording limits: [permission boundaries]
Output
Separate known facts, hypotheses and unknowns. Propose five open questions tied to the buyer's current workflow and next decision. Flag leading language, inferred budget or authority, and unsupported urgency.
Suggest one bounded next step and one respectful reason to pause.
Constraints
Use only the supplied material. Do not invent customer quotes, infer permission or claim a purchase is confirmed. Prepare a draft for human review. Do not send messages, access new private records or change records.Connect the next step to the relationship
Use Finta to keep relevant contacts, relationship context and working follow-ups organized. Ask Aurora to help prepare a brief or recap only from information it can access with permission, then check the draft against the conversation before using it. Do not label an AI-generated inference as a buyer commitment.
A successful discovery call can end with a smaller test, an introduction to the appropriate participant, a clear mismatch or a pause. Measure whether the next decision became clearer, not simply whether another meeting appeared on the calendar.
For the broader operating rhythm around these conversations, use the relationship-led sales playbook. Keep this call's facts and next decision visible within that rhythm rather than replacing them with a generic follow-up cadence.