Plenty of sales teams have bought AI tooling before deciding what it's for. The result is a stack of licences, a spike in generated emails and no measurable change in win rates. Starting from the tool almost guarantees that outcome.
Start where the work is hardest
Ask your sellers where their time goes and where they feel least prepared. Usually it's research before a first call, summarising what happened afterwards, and working out what to do next on a stalled deal. Those are the places AI earns its keep, because it removes friction rather than replacing judgement.
- Pre-call research: a short, accurate brief instead of twenty minutes of tab-hopping.
- Post-call capture: notes and next steps written while the detail is fresh.
- Deal review prep: an honest summary of what's missing before the manager asks.
- Practice: rehearsing a difficult conversation without burning a real meeting.
AI should reduce the number of things a seller has to hold in their head, not add to them.
Judgement doesn't transfer
AI can draft a message, summarise a call and suggest a next step. It cannot decide whether this deal deserves your attention this week, or read the pause after a difficult question. Teams that improve fastest use AI for preparation and admin, and keep the thinking with the seller.
Pick one workflow, measure the time it gives back, and only then widen it. Adoption follows usefulness.