Discovery Call Guide: Structure, 30 Questions, and a 10-Minute Prep Routine (With AI Call Prep)

Most discovery calls fail in the first five minutes, when the rep opens the slide deck. The prospect came to talk about their problem and got a tour of your features. Discovery is the opposite of a demo: its job is to learn enough to know whether a second meeting is worth anyone's time, and to earn that meeting by being the first vendor who listened. This guide gives you a five-part structure, thirty questions grouped by what they reveal, the metrics that show whether you actually listened, a ten-minute preparation routine, and a look at how AI call prep assembles the briefing so the ten minutes go on thinking rather than searching.

Quick answer: Structure a 30-minute discovery call as: a two-minute agenda and time check, fifteen minutes of questions about the prospect's situation, problem, impact and decision process, five minutes summarising what you heard, five minutes on fit and next steps, and a booked follow-up before you hang up. Ask 10 to 15 real questions, let the prospect talk more than you, and prepare by reviewing previous meetings, open commitments and CRM context. Finish with a same-day follow-up email.

Key takeaways

• Discovery is a listening exercise with a decision at the end, not a soft demo.
• Structure beats improvisation. Five parts, fixed timings, a booked next step.
• The best questions are follow-ups: "what happens then?", "how do you handle that today?", "who else feels this?"
• You cannot judge your own listening. Measure it: talk ratio, questions asked, monologue length.
• Preparation is the difference between generic questions and questions that only you could ask.

What a discovery call is for

A discovery call has three outputs. First, an understanding of the prospect's situation, the problem, what it costs them, who decides, and when. Second, a mutual judgement about fit: is this worth pursuing for both sides? Third, a next step, booked on the call.

Everything that does not serve those outputs is a distraction, including most of your product. Prospects on a discovery call do not want to see the dashboard; they want to feel understood. The sales methodologies that dominate B2B, from Neil Rackham's SPIN Selling through MEDDIC and its variants, differ in vocabulary but agree on the point: uncover the problem and its cost before proposing anything.

The five-part structure

Open

State the agenda and ask for permission: "I'd like to spend most of our time understanding what's going on for you, then tell you honestly whether we can help. Is that okay?" Then check the time: "We have until half past, does that still work?" Two sentences that set the tone for everything after.

Discover

Ask, listen, follow up. Take notes on their words, not your interpretation. Resist answering an objection that has not been raised. When they ask "so what does it do?", give a one-sentence answer and return to questions: "It handles exactly that, and I'll show you next time. Before that, can I ask how you deal with it today?"

Reflect

Play back what you heard in their words: "So if I've got this right, the main issue is X, it's costing you roughly Y, and the person who'd need to sign off is Z." Prospects correct you, which is the point. A reflected summary is also the raw material for your follow-up email.

Fit

Say what is true. If you can help, say how, in two sentences, tied to what they told you. If you cannot, say so; you will be remembered for it. If you are unsure, say what you need to find out.

Close

Propose the next step and book it on the call: "The right next step is a 45-minute session with your ops lead where we walk through your workflow. Here's my calendar, shall we pick a time now?" A booking link on the call converts far better than "I'll send times". The scheduling-link etiquette guide covers the phrasing.

30 discovery questions, grouped by what they uncover

Do not ask all thirty. Pick the ones that fit and go deep with follow-ups.

Situation: how things work today

1. Walk me through how you handle [the process] today.
2. Who is involved, and where does it start and end?
3. What tools are in the mix right now?
4. How long has it worked this way?
5. What changed recently that made this a priority?

Problem: what is not working

1. Where does it break down most often?
2. What have you already tried?
3. What happened when you tried that?
4. If you did nothing for six months, what would happen?
5. Who feels this most day to day?

Impact: what it costs

1. How much time does that take each week, roughly?
2. What does it stop you doing?
3. Has it cost you a customer, a hire, or a deadline?
4. How does your manager see this problem?
5. If this were solved, what would you do with the time?

Decision: who and how

1. Who else would be involved in a decision like this?
2. How have you bought tools like this before?
3. What would make this an easy yes internally?
4. What would make it a no?
5. Is there budget set aside, or would this need to be found?

Timing: when

1. When would you want this working by?
2. Is anything driving that date?
3. What happens if it slips?

Alternatives: what else is on the table

1. What other options are you looking at, including doing it in-house?
2. What do you like about them?
3. What would tip you toward one over another?

Fit and next step

1. Based on what we've discussed, what would you want to see next?
2. Who should be in that conversation?
3. What would you need from me before then?
4. Is there anything I haven't asked that I should have?

The question that most often changes the call is number 30. Ask it every time.

Follow-ups: where the real answers are

The first answer to any question is the rehearsed one. The useful answer comes from the follow-up.

• "What happens then?" turns a process description into a pain point.
• "How do you handle that today?" turns a complaint into a workflow you can map.
• "Who else feels this?" turns one person's frustration into an organisational problem.
• "Roughly how often?" turns "sometimes" into a number.
• "What did that cost?" turns a story into an impact.

A discovery call with ten questions and twenty follow-ups is better than one with thirty questions and none.

Measuring whether you listened

Reps consistently overestimate how much they listened. The only fix is to measure. Three conversational mechanics, all computable from a transcript, tell you how the call actually went.

Talk ratio. Your share of the speaking time. Most coaching models target 40 to 50 percent for the host, and discovery calls often run lower. PepoSmart's published call score awards full marks for a host share of 40 to 50 percent and reduces the score outside a 30 to 60 band; the talk-to-listen ratio guide explains the formula.

Questions asked. Ten to fifteen per 30 minutes is the benchmark PepoSmart uses. Fewer means you pitched; far more means you interrogated.

Longest monologue. The one four-minute stretch where you explained the roadmap is where the prospect checked their email. Keep any uninterrupted stretch under a minute on a discovery call.

Recorded calls make this automatic. The AI notetaker joins the Zoom, Google Meet or Microsoft Teams call, produces a transcript with speaker labels, and the coaching scorecard reports the ratio, question count, filler words and longest monologue for each call, with 30-day trends. Read the score alongside the recording, not instead of it: it measures mechanics, not whether the product fit.

The ten-minute preparation routine

Preparation is what separates "tell me about your business" from "I saw you opened a second warehouse in March; how has that changed the picking process?" Ten minutes is enough if they are spent well.

1. Two minutes: who are they? Role, tenure, what their team does, one recent thing the company announced.
2. Two minutes: what do we already know? Previous conversations with anyone at the company, what was promised, what was objected to.
3. Two minutes: the deal context. Stage, value, open objections, competitors mentioned, whether budget has come up.
4. Three minutes: three questions only you could ask. Specific to them, drawn from the first three steps.
5. One minute: define a good outcome. What next step would make this call a success?

The problem with this routine is that steps two and three require searching: old notes, the CRM, a colleague's memory. On a busy day they get skipped and the call opens generic.

How AI call prep assembles the briefing

This is the gap AI call prep closes. Before an upcoming meeting, PepoSmart assembles a prep sheet from what it already knows about the person, so steps two and three of the routine are done before you open the page. The AI call prep page describes it; here is what lands on the sheet.

Previous-meeting summaries. Recaps of recorded calls with this contact, each with its sentiment, call score and key topics. If a colleague spoke to them last quarter and the call was recorded, you know what was said.

Open action items. Commitments still owed to this contact from earlier meetings. Walking into a call not knowing you owe them a pricing sheet is avoidable.

CRM deal context. When HubSpot or Salesforce is connected on the Professional plan and above, the associated deal's name, stage, value, open objections, competitors and risk read appear on the sheet.

Coaching reminders. Personalised nudges from your own recent calls, such as "talk ratio has been high; ask more questions". The prep sheet is where the scorecard turns into behaviour.

Suggested questions. Three to five specific questions generated from the contact's actual history: their previous meetings, the CRM deal including open objections and competitors, any open action items, and your coaching focus. These are the questions only you could ask, drafted for you.

Two honest limits. A brand-new contact with no recorded history has almost nothing to brief you on, so the routine's first step, reading about them, is still yours. And the briefing is generated by AI from transcripts, so treat it as a prompt to check rather than a record of fact.

Common discovery call mistakes

Demoing. The prospect asks what the product does, and twenty minutes later they know your feature list and you know nothing. Answer in a sentence and return to questions.

Accepting the first answer. "We're looking to improve efficiency" is not a problem. Ask what that means, three times.

Asking about budget in the first five minutes. It reads as qualification for your benefit. Earn it with the impact questions first.

Solving the problem on the call. Once you start advising, the prospect stops disclosing. Save the solving for the next meeting.

Ending without a booked next step. "I'll send some times" is where deals go to die. Book it on the call.

Skipping the follow-up. The reflected summary, sent within hours, is how the prospect remembers you understood them. The follow-up email guide has templates and timing; if the call was recorded, PepoSmart drafts one from the transcript for you to review.

After the call

Three things, in order.

Send the follow-up the same day. Lead with what you promised, recap in three bullets, propose the next step with a booking link.

Log it. On the Professional plan and above, the recorded call's summary and approved action items sync to your CRM automatically, which the CRM sync guide describes. If you are logging by hand, the reflected summary is the note.

Tag the outcome later. When the deal closes or dies, tag the meeting as won or lost. Once you have a handful of each, the scorecard's win/loss correlation shows which discovery behaviours are associated with closing for your team, which is more useful than any generic benchmark.

A note on booking the discovery call itself

The best discovery call is the one that happens quickly. If your "Book a demo" link is really a discovery call, name it so ("30-minute discovery call") and describe what will happen, so prospects arrive expecting questions rather than a tour. For inbound volume, a round-robin event across the team keeps the first available slot close. For outbound, a one-off meeting with three hand-picked times in your email removes the scheduling exchange entirely.

Summary

Run discovery as a structured listening exercise: open with an agenda, spend half the call on questions about situation, problem, impact and decision, reflect what you heard, be honest about fit, and book the next step before you hang up. Ask ten to fifteen real questions, follow up relentlessly, and measure your talk ratio rather than trusting your memory of it. Prepare in ten minutes, and let AI call prep do the searching so the minutes go on the three questions only you could ask. PepoSmart's call prep sheet, coaching scorecard and follow-up drafts are all built from the recorded call, on Starter and above with two recordings to try on the free plan; the sales solution page shows the full flow and the pricing page has current allowances.