In a normal day, reps use AI in the gaps around selling: researching an account before a call, turning the call into CRM notes and a follow-up the prospect gets the same afternoon, replying to new leads in minutes, and drafting a calm answer to a discount request that stays inside the pricing rules. The single habit that pays off fastest is the same-day follow-up, which you can practice in Send a Same-Day Follow-up After Every Discovery Call using ChatGPT.
What AI does not do is sell. It does not build trust on a call, read the room or decide what to offer. It saves the hour of admin that usually follows each conversation, so more of the day goes to the conversations themselves.
What does an AI-assisted day look like?
8:30, before the first call. Paste the prospect's website "About" page and your notes from the last touch into ChatGPT or Claude and ask for five questions to ask and two likely objections. Check anything it states as fact about the company.
9:00, overnight leads. If your team has an automation, every website lead already got a personal first reply within minutes and the hot ones were posted to the team channel. You start with the ones flagged hot.
10:00, discovery call. Your call recorder (Gong, or the transcript from Zoom, Google Meet or Teams) captures the call. Some reps use a live assistant like OLVA to keep track of questions during the call.
10:45, after the call. Paste the transcript into AI and ask for a summary, decisions, next steps with owners and dates, open questions and deal notes for the CRM (budget, decision maker, need and timeline) from only what the prospect said. Send the follow-up before lunch.
1:30, CRM update. Paste the deal notes into HubSpot or your CRM. Many CRMs now have built-in AI that can summarize records and draft emails as well.
3:00, the hard email. A prospect asks for a big discount. AI drafts a reply from the thread and your discount rules, then checks every promise against those rules before you send.
4:30, tomorrow. Ask for a short list of open deals with no next step dated, from your CRM export.
What does this look like in a real lesson?
In the discovery call lesson you are Jordan, a rep at a payroll software company, just off a call with an architecture firm that wants to switch payroll providers by January 1. You copy the Google Meet transcript, paste it into ChatGPT under a request for five parts, including deal notes taken only from what the prospect said. The next steps table comes back with two due dates marked Not stated instead of guessed. You fill them in yourself (pricing by tomorrow, the timesheet export from the prospect by Monday), then ask for a short follow-up email with no price quoted and no promise about the tax forms.
Two more lessons fill out the day. Speed-to-Lead Agent builds an n8n workflow that scores each web lead from 1 to 10, emails a personal first reply right away and pings Slack only when the score is 8 or more. Hold Your Price When a Prospect Asks for a Discount catches a draft that agreed to a discount and a free month only the sales director could approve.
What mistakes should I avoid?
- Promising what you cannot approve. AI drafts often say yes to discounts, dates and features. Check every promise against your rules.
- CRM notes built on guesses. Tell AI to use only what the prospect said, so the forecast is not wishful thinking.
- Sending without reading. Names, prices and dates go wrong in drafts. Read before you send.
- Recording without telling people. Some states require everyone on a call to agree. Say you are recording or taking AI notes.
- Bulk AI emails that ignore the rules. Commercial email still has to follow CAN-SPAM, and texts and automated calls have their own rules. See Can AI write cold outreach that does not get ignored or flagged as spam?
Which tools do reps use most?
- ChatGPT or Claude for call prep, follow-ups, objection replies and CRM notes.
- HubSpot or your CRM, increasingly with AI that writes emails and summarizes records.
- Gong to record and analyze sales calls.
- OLVA as a private assistant during live calls.
For the full lesson path, see the sales rep start guide. For automations a team can share, see AI automation for sales teams, and for copy-ready prompts, AI prompts for sales reps.
Key takeaways
- AI fits around the selling: prep, follow-up, CRM notes, lead replies, hard emails.
- Same-day follow-ups from the transcript are the fastest win.
- Deal notes come only from what the prospect said, with Not stated for gaps.
- Check every discount, date and promise before it goes out.
Frequently asked questions
- Is it OK to paste a call transcript into ChatGPT?
On your company's approved plan, usually yes, but follow your rules on confidential deal terms. Leave out anything the prospect shared in confidence that the summary does not need.
- Can AI update my CRM automatically?
Yes, with automations or built-in CRM AI, but start by pasting notes yourself so you see what it gets wrong. Automate once the notes are reliably accurate.
- Will AI make every rep's emails sound the same?
It will if everyone uses one-line prompts. Give it your own past emails as examples and edit the draft in your own words.