What Are the Best AI Prompts for Sales Teams? 9 to Copy and Use

The best AI prompts for sales teams do the unglamorous work: researching an account, planning a discovery call, writing the follow-up and checking the pipeline for wishful thinking. They don't write your pitch for you. They make the rep better prepared before every conversation and quicker after it.

I've trained a lot of sales teams to use ChatGPT, Claude, Microsoft Copilot and Gemini, and the pattern is always the same. The prompts that stick are tied to a moment in the sales cycle that already exists. Here are nine I'd hand to any UK sales team, and how to roll them out safely.

What are the best AI prompts for sales teams?

The best AI prompts for sales teams give the AI a role, the real material (call notes, CRM export, buyer emails), a clear output and firm limits. If you've read my guide to how to write prompts for AI, this is that method applied to selling. Here are the nine prompts on this page and where each fits.

Swap the [square brackets] for your details. And don't paste confidential deal terms or customer personal data into an AI tool your company hasn't approved.

Which ChatGPT prompts help with account research and call prep?

The most useful ChatGPT prompts for sales prep turn scattered research into a one-page brief and a call plan the rep can glance at. Prep is where AI pays back fastest, because busy reps skip it.

Prompt 1: Account research

You are a B2B sales researcher helping me prepare for a first meeting with [company name], a [sector] business based in [town or region].

Using only the material I paste below (website copy, annual report summary, recent news), write a one-page account brief with these headings:
1. What they do and who they sell to
2. Likely priorities this year
3. Pain points our [product or service] could address
4. Who usually owns this decision in a firm like theirs
5. Three questions I should ask

Mark anything you are inferring rather than reading directly as (inferred). Do not invent figures, names or dates.

Material:
[paste here]

Why this prompt works

It limits the AI to the material you supply and makes it label its guesses, so you can see what is evidence and what is speculation.

If your tool has web search switched on, you can ask it to find the material itself, but check every claim. AI research is confident and sometimes wrong, so read my notes on how to judge the accuracy of AI information before a rep repeats a "fact" to a prospect.

Prompt 2: Discovery call preparation

Act as an experienced B2B sales coach. I have a 30-minute discovery call with a [job title] at [company]. What I know so far: [two or three lines of CRM notes]. Our offer in one line: [description].

Prepare a one-page call plan with:
- A two-sentence opening that sets an agenda and earns the right to ask questions
- Eight open questions in order: situation, problem, impact, decision process, timing
- Two follow-up questions for when they give short answers
- The one thing I must not leave the call without

Plain British English. Nothing I would be embarrassed to read aloud.

Why this prompt works

The question order mirrors how good discovery actually flows, and the last line forces a clear objective for the call.

What are good AI prompts for objection handling and proposals?

Good AI prompts for objection handling make the AI play a sceptical buyer so the rep can practise, and good proposal prompts structure the client's own words into a draft.

Prompt 3: Objection handling practice

I want to practise handling objections. Play a [job title] at a [size] [sector] company who is interested but sceptical about [our product or service].

Raise one realistic objection at a time: first price, then timing, then "we already have a supplier". After each of my replies, stay in character and respond as this buyer would. Then step out of character in one short paragraph headed Coach: telling me what worked, what didn't, and a stronger line I could have used.

Do not let me off lightly. Stop after five rounds and give me a score out of 10 with reasons.

Why this prompt works

Rehearsal beats a script. The coach notes turn a role play into feedback the rep can use on the next real call.

Make it better

Paste in two or three real objections from recent calls. Your actual market beats a generic one.

Prompt 4: Proposal executive summary

Write a proposal for our new client.

You are a bid writer for a UK [sector] firm. Draft the executive summary and scope for a proposal to [client name], using the call notes and requirements below.

Structure:
- Their situation, in their words (two short paragraphs)
- The outcome they want
- Our recommended approach in three to five steps
- What is included and what is not
- Next step, with [date] as a placeholder

Under 500 words, plain British English, no superlatives. Do not include prices. Where the notes leave a gap, write [CHECK: question] instead of guessing.

Notes:
[paste here]

The [CHECK: question] instruction is the part I'd keep if you forget everything else. It stops the AI filling gaps with plausible fiction. For more on this, see my guide to AI for business emails and proposals.

Which AI prompts write better sales follow-up emails?

AI prompts write better follow-up emails when they include what the buyer actually said and ban the generic opener. A specific follow-up sent within the hour beats a polished one sent on Thursday.

Prompt 5: The meeting follow-up

Write a follow-up email to [first name], [job title] at [company], after today's meeting.

What they raised: [bullet points]
What we agreed: [actions and owners]
Proposed next step: [date and format]

Rules: under 150 words. Subject line of eight words or fewer. Open with something specific from the conversation, not "Great to speak today". Restate their problem before mentioning our solution. One clear call to action.

Give me two versions: one warm and conversational, one brisk for a busy senior buyer.

Why this prompt works

The rules ban the generic opener that every buyer skims past, and the two versions let the rep match the person.

In Microsoft Copilot, if your organisation has it switched on for Teams and Outlook, you can point this prompt at the meeting recap rather than pasting notes. My list of Copilot prompts for Outlook and Teams covers that workflow in more detail.

Prompt 6: The stalled deal

A deal has gone quiet. Here is the timeline from our CRM: [dates, meetings, stage, last contact]. The buyer's last message said: [paste].

Act as a sales manager reviewing this deal.
1. List three likely reasons it has stalled, ranked by how well the evidence supports each.
2. Draft a re-engagement email under 100 words that makes it easy for them to say yes, no or not now. No fake urgency, no "just checking in".
3. Suggest one next action that isn't an email.

Why this prompt works

Diagnosis before drafting stops the rep sending the same chaser for the fourth time.

How can sales leaders use AI prompts for pipeline review and win/loss analysis?

Sales leaders should use AI prompts to audit pipeline data and find patterns in won and lost deals, not to predict next quarter's number. It spots inconsistencies a tired manager misses. It is not a forecasting model.

Prompt 7: Pipeline review

You are a sales director preparing for Monday's pipeline review. Below is an export of open opportunities with these columns: deal ID, stage, value, close date, last activity date, next step, owner initials.

Flag:
- Close dates already in the past
- No activity for more than [14] days
- Late-stage deals with no next step
- Anything where stage or value looks out of line with the activity

Show a table of flagged deals with the reason and one question I should ask the owner. Then give a three-line summary of pipeline health. Do not forecast revenue.

Data:
[paste export]

Why this prompt works

It turns the AI into a ruthless auditor of the data rather than a fortune teller, which is the job it does well.

This works well in Copilot in Excel against a CRM export, which keeps the data inside your Microsoft 365 environment. There are more ideas in my Copilot prompts for Excel.

Prompt 8: Win/loss analysis

I am analysing why we win and lose deals. Below are anonymised notes from [number] closed deals in the last [period], each marked WON or LOST with the reason the rep recorded and any buyer feedback.

1. Group the reasons into themes.
2. Show a table: theme, wins, losses, one example quote.
3. Tell me the three patterns most worth acting on, and what evidence would prove or disprove each.

Keep what buyers said separate from what reps said. Say plainly where the sample is too small to conclude anything.

Notes:
[paste here]

Why this prompt works

Separating buyer voice from rep voice exposes the comfortable stories a team tells itself about lost deals.

Prompt 9: Sales coaching

Act as a sales coach supporting a team member's development. Below is an anonymised transcript or set of notes from a discovery call.

Assess it against five areas: agenda setting, quality of questions, listening (did they follow up on what the buyer said?), handling the main objection, and agreeing a next step.

For each area give a rating (strong, OK or needs work), one short quote as evidence and one specific thing to try next time. End with the single biggest improvement.

Tone: direct but encouraging. I will share this in a 1:1.

Transcript:
[paste here]

Why this prompt works

Evidence-based ratings make coaching conversations about the call, not about the person.

How should a sales team use AI prompts safely?

A sales team should use AI prompts safely by sticking to approved tools, stripping out personal data and checking every fact before it reaches a customer. Sales teams handle more personal data than most, and a contact record in your CRM is personal data under UK GDPR.

Why are AI prompts alone not enough for a sales team?

Prompts alone aren't enough because a sales team needs shared habits, not a document of clever text that three people open once. What changes behaviour is practice on real deals, agreed rules and a manager who uses the prompts visibly.

Here's the rollout order I suggest to sales directors.

1. Pick three moments.Choose the three points in your sales cycle where reps lose the most time. For most teams that's prep, follow-up and pipeline hygiene.
2. Agree the rules.Decide which tool is approved and what data can go in it. Write it down on one page.
3. Practise on live deals.Run a session where every rep uses the prompts on a deal they're working this week, not a made-up example.
4. Share what worked.Keep the best versions in a shared team library and let reps improve them.
5. Review in a month.Ask what's being used, drop what isn't and add the next moment.

This is the work we do in our training. Basic AI Training gets the whole team confident with ChatGPT, Claude and Copilot, and the Microsoft Copilot Deep Dive suits teams living in Outlook, Teams and Excel. If you want to see the wider case for sales, read how to win more business with AI. For prompts covering the rest of the business, start with the best AI prompts for business, and if you're planning a rollout, the best AI prompts for team AI training is the next page in the series.

Get your sales team prompting on real deals

We train UK sales teams in person and online to use ChatGPT, Claude and Copilot on the deals they're working this week, with clear rules on customer data. Tell us about your team and we'll suggest a starting point.

See all programmesTalk To Harry

• Basic AI Training Your whole team, prompting properly in a day.
• Prompts, Context and Skills Deep Dive A half day on Claude and ChatGPT.
• AI Capability Assessment See where your team stands first.

Harry Lang
Harry Lang founded The Oxford AI School in Witney, Oxfordshire. He trains UK businesses and teams to use AI well, writes the Marketing Unfiltered newsletter, and is the author of Learn To Fly With AI.