Guide
How to prepare for a sales call with ChatGPT
To prepare for a sales call with ChatGPT, paste what you know about the prospect, with where each fact came from, and ask for a one-page plan. A three-line snapshot of the company. Two or three hypotheses about their problem that the call can confirm or kill. Eight questions, ranked, to spread through the call: Gong found 11 to 14 targeted questions had the best odds, so eight plus follow-ups lands there. Four or five yes-or-no items you must learn to qualify the deal, a few red flags, and the opening and close. Keep it under 350 words, and check every fact before you dial.
Short answer
- Snapshot: three lines, each fact with its source.
- Hypotheses: two or three, each with a reason from the research.
- Eight questions, ranked, spread through the call.
- Qualify: pain, a number, who signs, a date, what they use now.
The brief to copy
Call prep brief
"Prepare a one-page plan for a discovery call with [name, role] at [company]. We sell [product] to [type of company].
Snapshot in three lines: what they do, size or stage, and the one fact most relevant to what we sell. Cite where each fact came from.
Two or three hypotheses about their problem, each with the reason from the research. Ranked.
Eight open questions, ranked: two to confirm the top hypothesis, two on impact, one on what they've tried, one on timing or the trigger, one on how they decide and who signs, and one open closer ('What would make this call worth it for you?'). Spread them through the call.
Four or five yes-or-no items to qualify: pain in their words, a cost or number, who signs, a date or trigger, what they use now. Three to five red flags. One opening line and the next step to propose.
Under 350 words. Research: [paste]"
A worked example
Illustration: a made-up business. Dev sells inventory software to wholesalers and has a call with the operations manager of a kitchenware distributor that, according to its careers page, is hiring for a second warehouse. From the plan:
From the call plan
Snapshot: Kitchenware wholesaler, about 60 staff (LinkedIn). Hiring warehouse staff for a second site (careers page, 2 November). Uses spreadsheets for stock, per a 2025 job ad.
Top hypothesis: with two sites, stock counts are manual and slow, so orders ship short. Reason: second site plus spreadsheet stock.
First two questions: "How do you know today what's in stock at each site?" · "When did an order last ship short, and what happened?"
Qualify (yes/no): problem in their words · a cost or number · who signs · a date (the second site's opening) · what they use now.
Red flags: a stock system already chosen for the new site; nobody owns the problem; "just looking" with no date.
Close if qualified: a 30-minute walk-through with the ops manager and whoever signs, before the second site opens.
Dev checks each fact on the pages it came from and drops the spreadsheet line when the 2025 ad turns out to be for a different company with the same name.
Check before the call
- Every fact has a source you opened yourself.
- Every hypothesis has a reason from the research.
- Questions are open and spread across the call.
- You know the minimum you need to book a next step.
- Nothing confidential went into a personal account; see what not to paste.
Next step: keep the brief
You'll prep calls every week from the same brief. Sales prompts, saved shows where to keep it so you only paste the research. After the call, see turning call notes into CRM notes.
Skills for this
- SalesDiscovery call prepBuilds a one-page discovery call prep from the prospect's website, LinkedIn and notes: hypotheses, 8 ranked questions, what to qualify, red flags.
- SalesAccount expansion planBuilds an expansion plan for an existing customer: renewal health, stakeholder map, current use, whitespace, risks, and 90-day plays with owners.
- SalesObjection handlingDrafts replies to price, timing, competitor, security and "send me info" objections: acknowledge, clarify, respond, confirm. Never discounts first.
Questions
How do I prepare for a sales call with ChatGPT?
Paste what you know about the prospect, with where each fact came from, and ask for a one-page plan: a three-line snapshot, two or three hypotheses about their problem, eight ranked questions, four or five yes-or-no items that qualify the deal, red flags, and your opening and close.
How many questions should I ask on a discovery call?
Gong's analysis found the best odds with 11 to 14 targeted questions, and success fell back to average beyond that. Prepare about eight and let follow-ups make up the rest, spread through the call rather than all at the start.
What should I learn on a discovery call?
Enough to tick four or five yes-or-no items: the problem confirmed in their words, a cost or number, who signs, a date or trigger, and what they use today. Those map to budget, authority, need and timing.
AI skill
Have your AI prepare your next discovery call
Put this skill in AI Skill Pod and everyone on your team gets it in their Claude, Cursor or ChatGPT (Pro, Business, Enterprise): one invite link, free to start. With Discovery call prep, each rep's AI builds the call plan from the research: a sourced snapshot, hypotheses to confirm or kill, eight ranked questions, the yes-or-no items that qualify the deal, red flags, and the opening and close, on one page.
14 days free · No card needed
Sources
- Gong: nailing your sales discovery calls — 11 to 14 targeted questions had the best odds, more fell back to average; top reps spread questions evenly through the call, average reps front-load them; correlation, not causation (read 2026-11-05)
- Salesforce: BANT vs MEDDIC — budget, authority, need, timing (read 2026-11-04)
- OpenAI Help: Data controls — personal plans train on chats by default unless turned off; read through search-result text (read 2026-11-05)