How can sales teams use AI tools like ChatGPT, Copilot and Claude?

The best use of AI in sales is clearing the admin so your salespeople spend more time with buyers. ChatGPT, Copilot and Claude can research an account, prepare a call, draft the follow-up, tidy CRM notes and write a first-draft proposal. They can't own the relationship, set the price or make the promise. Give your team an approved tool, firm rules on customer data and the habit of checking every line, and those hours go back into selling.

Here's where AI earns its place in a sales week, what it should never be left to do and what good training covers.

Why does a sales team need its own AI rules?

Because almost everything a salesperson types is about a customer. Names, phone numbers, deal values, what someone said on a difficult call. That makes sales one of the teams where a careless paste does the most damage.

It's also a team that works against the clock. When a proposal is due by five, the quickest tool to hand wins, and if nobody has shown your salespeople the approved way to work, that's a free app on their phone. Clear rules and proper training fix that, because the approved tool becomes the fast one.

Where does AI save a sales team time?

Start with account research. Before a first call, a salesperson can paste in a prospect's own website copy or annual report and ask for a plain summary of what the business does, who it sells to and what's changed recently, then for the five questions most worth raising. Anything the tool says about the company still gets checked against the company's own website or Companies House, because a confident summary can be out of date or simply wrong.

Call prep is next. Give the tool your own notes from the last conversation and ask for a one-page brief: who you're meeting, what was agreed, what's still open and which objections to expect. Claude and ChatGPT both make good sparring partners here. Ask one to play a sceptical finance director and rehearse your answer on price before you're in the room.

Follow-up emails are a quick win. After a call, jot down three or four points about what was said and agreed, and ask for a follow-up in your own tone that recaps the next steps and dates. If your team works in Microsoft 365, Copilot in Outlook summarises long email threads, and Microsoft's support page on Copilot in Teams meetings, checked in October 2026, says it can summarise key discussion points, "including who said what", and suggest action items. The meeting has to be transcribed, unless the organiser has switched Copilot on without transcription.

CRM notes follow the same pattern. Rough notes typed between meetings become a clean, structured record with a clear next step and a date. Your pipeline stays current because updating it stops being a chore.

Proposals come last. Give the tool your proposal template, the buyer's stated priorities and the case studies you're allowed to use, and it'll draft a summary written around what this buyer cares about. Your salesperson then owns every line, and every price and term gets checked against your own price list.

What should never be left to the AI?

Anything that becomes a promise. AI tools write fluently even when they're wrong, and in sales a wrong sentence turns into a commitment the moment a customer reads it: a feature you don't offer, a discount nobody signed off, a delivery date nobody agreed. So the rule is short. Nothing goes to a customer until the salesperson sending it has read every word and stands behind it.

Pricing and terms work the same way. AI can set out the case for and against a discount in seconds, but the decision, and the accountability for it, sits with whoever owns the deal. So does the choice of which prospects to chase and what goes into a contract.

How do you keep customer data safe with AI?

What follows is general information, not legal advice, and your data protection lead sets your actual policy.

CRM notes about named contacts are personal data, and UK GDPR's accuracy principle applies to them. The ICO's guidance on accuracy, dated 2025, says "You should take all reasonable steps to ensure the personal data you hold is not incorrect or misleading as to any matter of fact." An AI summary that gets a contact's job title wrong, or records a commitment they never made, puts exactly that kind of error into your CRM. Whoever saves the note checks it.

Customer details also never go into an AI tool your organisation hasn't approved, because business plans and personal accounts come with different data terms. Anthropic's privacy centre, updated in March 2026, says that for its commercial products, including Claude for Work Team and Enterprise plans and the API, "We will not use your chats or coding sessions to train our models" unless you choose to join its Development Partner Program. Ask your IT lead or supplier which terms cover the licences you hold. For your salespeople, keep it to one line: the approved tool, signed in with a work account, and nothing else.

Which AI tool should a sales team use?

The one your organisation has approved. If your business runs on Microsoft 365, Copilot works inside Outlook and Teams, where your team already handles email and meetings. If you've approved ChatGPT or Claude, both handle drafting, role play and working through material you give them. Our article on which AI tool your team should learn first goes through that choice in detail.

Once you've chosen, consistency matters more than the logo: everyone using the same tool in the same safe way, and your best salesperson's prompts shared with the team instead of living in one person's head.

What should AI training for a sales team cover?

Good training is built on your team's real work: your proposal template, your CRM fields, your products and the objections your salespeople hear every week, with made-up or anonymised customer details in every exercise. A session should cover:

• Prompts for the jobs that fill the week: account briefs, follow-ups, CRM notes and proposal sections.
• What must never go into any AI tool, and how to strip customer details out before you paste.
• How to check outputs for wrong facts, prices, names and promises.
• Objection and pricing role plays with the AI as the buyer.
• A shared set of prompts the team keeps using after the session.

Fit the training around your sales calendar, away from quarter-end and any big push, when nobody's mind is on a workshop. Sessions run as a half day, a full day or a short series, and a series gives salespeople time to try what they've learned on live deals and bring back what worked.

Our AI training for sales teams uses exactly these tasks. It's private and hands-on for up to 12 people, at your premises anywhere in the UK, at a venue or live online, with one fixed price per group confirmed before you book. Attendees rate our workshops 5/5. DLR Media has signed the Responsible AI Use Campaign's 10 principles, and responsible use is built in throughout. See how bookings work on our private training page, or how we train other departments on our AI training for staff page.

Frequently asked questions

Can ChatGPT write our sales emails?

Yes, as a first draft. Give it your own notes and tone, then a salesperson checks every fact, price and promise before it's sent.

Can we put customer details into an AI tool?

Yes, but only into a tool your organisation has approved for customer data, used with a work account and within your policy. A free personal account is never the place for them.

Will AI replace our salespeople?

No. It drafts and prepares. The conversations, the pricing and the accountability for every promise stay with your salespeople.

Can AI keep our CRM up to date?

It can turn rough notes into a clean, structured record in seconds. A salesperson still checks every name, date and commitment before it's saved, because a CRM is only useful if it's right.

Plan your sales team's AI training

Tell us which AI tools your team has and what fills their week, and we'll build the session around it. Call +44 1382 699595 or plan your sales team's training.