Account research and first-draft outreach are where most sales teams start with AI, for good reasons: the task is clear, the time cost is high, and a human can check the output before it goes out.
What AI does well here
- Summarizes an account: annual reports, news, job posts, tech stack, past CRM history.
- Finds the angle: links a trigger event (new leader, funding, expansion) to your offer.
- Drafts a first email or call opener the seller then edits.
- Prepares a meeting brief: who is attending, what they care about, open questions.
What the evidence says
- LinkedIn's 2025 research reports that 38% of sellers who use AI to research leads and companies save more than 1.5 hours a week, and that sellers who improved response rates with AI saw an average lift of 28%.
- Salesforce's State of Sales 2026 says sellers expect agents, once fully implemented, to cut prospect research time by 34% and email drafting by 36%.
- Microsoft reported that Lumen Technologies estimated Copilot saves its sellers an average of four hours a week, which Lumen equated to $50 million annually.
These are self-reported or vendor-published figures; use them as a reason to test, and measure your own.
Where it goes wrong
- Wrong facts: the model "remembers" a product, a title or a figure that isn't true. Sellers must check before sending.
- Generic personalisation: "I saw your company is growing" at scale is still spam.
- Time saved but not reinvested: if research is faster but meetings booked don't rise, nothing changed for revenue.
How to run it
Give sellers one approved tool inside the CRM or inbox, a short prompt guide for your ideal customer profile, and a rule: AI drafts, humans send. Measure prep time, positive reply rate and meetings held against a comparable team.