What AI actually looks like for a sales team
Not "AI will write your emails." What sales teams are genuinely using Claude for, what works, and the one thing most reps get wrong.
Sales is one of the best fits for AI — high-repetition tasks, clear quality signals, and enough volume that even a small time saving per rep adds up fast. Here's what it actually looks like when sales teams use Claude well.
What works
Pre-call research. This is the highest-value use case. Before a call, a rep pastes in a prospect's company name, LinkedIn summary, and any recent news. Claude produces a structured briefing: likely pain points, relevant company context, questions worth asking, potential objections. What used to take 25–30 minutes takes 3–5.
Setup: a Claude Project with your ICP description, your product's differentiators, and your most common customer situations. Claude then tailors the briefing to your product context, not just generic research.
Follow-up email drafts. After a call, paste your rough notes. Claude drafts the follow-up referencing specific things that came up. The rep edits and sends. Handle time drops significantly, and the emails are more specific than the average template.
Personalising outbound sequences. Generic outbound performs poorly. Claude can take a prospect's job title, company stage, and recent news and produce a personalised first line for each email in a sequence. You're still using templates for structure — Claude personalises the hook.
Summarising long email threads before re-engaging. A prospect went cold three months ago. You want to re-engage. Before writing anything, paste the full thread. Claude summarises: where the conversation stopped, what was agreed, what objections came up. You respond from a position of context, not guesswork.
Objection prep. Before a call with a prospect who's been pushing back on pricing, paste their objections and your notes. Claude helps you think through responses, stress-test your positioning, and identify the real concern underneath the stated objection.
What doesn't work
Fully automated outbound. Teams that have tried fully AI-generated cold outreach report that response rates often drop — it's detectable. Use Claude to personalise and assist, not to automate entirely.
Using Claude without your product context. Generic Claude giving generic answers about your product is worse than nothing — it produces confident-sounding outputs that are subtly wrong. Your Claude Project needs your product one-pager, pricing structure, and ICP description loaded in before anyone on the sales team uses it for customer-facing work.
Replacing call prep with AI briefings. The briefing is a starting point, not a substitute for knowing your prospect. Reps who read the briefing and stop there produce conversations that feel like they're following a script. Use it to supplement your judgment, not replace it.
The setup that takes 30 minutes and pays off immediately
- Create a Claude Project named something like "Sales — [Your Company]"
- Add your product description, key differentiators, and common objections to the Project instructions
- Add your ICP description: company size, role, typical pain points, what they usually compare you to
- Write a simple prompt template your team uses before every call: "Here's the prospect. Here's what I know. Give me a 5-point briefing."
- Share the Project with the team
From that point, every rep has the same starting point. The quality of briefings stops varying by how good each rep is at research.
The number to track
Time from meeting booked to call-ready. If your team is spending more than 20 minutes prepping for a discovery call, Claude should cut that to under 10. If it's not, your Project instructions need work.