LearnGrokPrepare one call document before each first meeting. It should give you a short question set, the evidence behind each question, and a clear next-step decision after the call.
This method is for sales representatives who have a stated prospect problem, a few account signals, and missing qualification details. It stops you asking a generic list of discovery questions that tells you little about whether the opportunity should move forward.
Key point
Start with gaps, not a stock question list
Every question should either test the stated problem, explain an account signal, or fill a qualification gap.
1. Build the input packet
Create a working note called First discovery prep. Use four sections. Keep each item factual and label anything you do not know as Unknown.
- Stated problem: Copy the prospect's own words from an email, form, referral note, or previous conversation. Include the source.
- Account signals: Record observable changes that may matter to the call. For example: a new leader, a team expansion, a public product change, a job advert, or an announced initiative. Add where you saw it and the date you recorded it.
- Qualification gaps: List the fields your team needs before moving an opportunity ahead. Typical fields are business problem, affected team, current process, business impact, decision process, target timing, and next agreed action.
- Known constraints: Add what is already confirmed, such as the meeting length, attendees, existing supplier, or a stated deadline.
Do not turn a signal into a fact. A job advert may suggest growth. It does not prove that the prospect has an active project or a problem your offer can solve.
Watch out
Keep evidence and assumptions separate
Write Signal: three open operations roles rather than The operations team is understaffed. The first is a fact you can ask about. The second leads the witness.
2. Ask for questions in four groups
Give the model your input packet and ask it to produce questions only from that material. Use this prompt, replacing the bracketed text:
Create a first discovery-call question set from the material below.
Stated problem:
[Paste the prospect's words and source]
Account signals:
[Paste signals, sources and dates]
Qualification gaps:
[Paste the missing fields]
Known constraints:
[Paste attendees, meeting length and confirmed details]
Group the questions under: business impact, current process, decision process and urgency.
For each question, state which fact or qualification gap it addresses.
Use open questions first. Add a short follow-up question only where it would clarify an answer.
Do not state or imply facts that are not in the material. Mark unsupported assumptions as `Do not use`.
Return no more than three primary questions in each group.
The model is helping you draft. You remain responsible for the account facts, the wording you use, and the decision you record. Product behaviour and available features can vary, so check the current documentation when your workflow depends on a particular capability: xAI documentation overview.
3. Edit the questions before the meeting
Read every proposed question against the input packet. Delete anything that does not have a job. A good first meeting rarely needs every possible qualification question.
Arrange the four groups in the order below. This order earns context before you ask the prospect to explain internal decisions.
- Business impact: Establish who is affected, what outcome is at risk, and why the issue matters.
- Current process: Learn what the team does now, where work slows down, and what has already been tried.
- Decision process: Identify who experiences the problem, who evaluates options, who approves a change, and what must happen before a decision.
- Urgency: Test whether there is a real timing event, consequence of delay, or agreed next milestone.
Use the stated problem to write the first question in the prospect's language. If their note says reporting is slow, ask: Where does the reporting process create the most delay for the team? Do not begin with a product-shaped question such as Would automation solve your reporting problem?
Check
Test each question before you keep it
You should be able to point to the exact gap it fills or the exact evidence that prompted it. If you cannot, remove it.
4. Create the call document
Put the final questions and the decision fields in one document. Do not keep questions in one place and notes in another. You need to compare what you planned to learn with what the prospect actually said.
Use this structure:
| Section | Before the call | During or after the call |
|---|---|---|
| Business impact | Two or three questions | Affected team, consequence, desired outcome, exact prospect wording |
| Current process | Two or three questions | Current steps, tools or suppliers mentioned, failed attempts, friction point |
| Decision process | Two or three questions | Participants, evaluation route, approval point, missing stakeholder |
| Urgency | Two or three questions | Timing event, cost of waiting as described, next milestone |
| Decision | Draft decision rule | Advance, nurture, or disqualify, with evidence and owner of next action |
Leave a Not discussed option beside every field. Empty fields look like negative answers when they may simply mean you ran out of time.
Write a provisional decision rule before the call:
- Advance when the prospect confirms a meaningful problem, a current process worth changing, a plausible route to a decision, and a time-bound next action.
- Nurture when the problem appears real but timing, ownership, or a decision route is not yet clear. Record the event that would justify reopening the conversation.
- Disqualify when the stated problem is outside your team's fit, there is no credible reason to change, or the prospect confirms that the opportunity should not continue. Record the prospect's reason without rewriting it into a softer story.
These are working sales decisions, not predictions. Keep the evidence beside the decision so a manager or colleague can see why you made it.
5. Run the meeting from the document
Open with the prospect's stated problem and ask for correction: You mentioned [problem]. Is that still the right place to start? This gives them control over the premise and may reveal that the situation has changed.
Ask one primary question at a time. Take notes in the matching field. When an answer is vague, use the prepared follow-up rather than moving to the next section. For example, if they say the issue affects several teams, ask which team feels the delay first and what happens next.
Do not force all four sections into a short meeting. Get enough evidence to make the next decision, then agree a specific action. The action should name the person, meeting or material needed, and the date or triggering event.
6. Check whether the output is misleading
The question set is wrong if it assumes a problem, asks the prospect to confirm your research, or produces an advance decision with no supporting notes. Review these failure signs after the call:
| If you see this | It usually means | Do this next |
|---|---|---|
Several questions start with I saw that you are... |
You treated a signal as proof | Reframe as an open question about the current situation |
| Notes contain opinions but no prospect wording | You inferred impact or urgency | Add the exact quote, or mark the field unknown |
| The decision process is blank | You discovered pain but not how change happens | Nurture or schedule a focused follow-up with the relevant stakeholder |
The next step is send information |
There is no testable mutual action | Agree what the prospect will review, who joins next, and what decision the next meeting supports |
Check
A completed document has evidence, not just ticks
An Advance decision should be readable by someone who did not attend the call. They should see what was confirmed, what remains unknown, and why the next action is justified.
When it does not work
If the research is thin, do not ask the model to fill the gaps. Build a smaller set around the stated problem and mark the rest as unknown. If the output is generic, tighten the input packet and require each question to cite a specific gap or signal. If the prospect changes the subject in the meeting, update the call document immediately, then make the decision from the new evidence rather than the original prep.
