Where AI earns its keep in a sales team

Most of what gets marketed to sales teams as AI falls under the heading of prediction: lead scoring, intent signals, a number beside every deal telling you how likely it is to close. Some of this is genuinely useful, and we sometimes use it. In my experience, though, the returns are smaller and harder to defend than the pitch suggests, largely because the predictions are only ever as good as the records underneath them. Where I see the money actually being made is in less glamorous work that improves those records, and it earns its keep quietly, every week, rather than in a demonstration.

The first of these is CRM hygiene. In most teams I visit, the pipeline is a record of what people remembered to type, not what actually happened. A call takes place, something important is said, and the CRM entry depends on whether somebody had the time and the inclination to write it up afterwards. AI can sit in the background here, turning calls and email threads into structured updates while people get on with their jobs. The record gradually starts to reflect reality, and everything that depends on the record, forecasting most of all, becomes more trustworthy as a result.

The second is meeting notes that stay. A good summary, filed with the commitments pulled out of it, within minutes of the conversation ending, changes the texture of the working week. Nobody spends their evening on write-ups. The next person to touch the account starts from what was actually said rather than from a colleague's recollection, and knowledge stops living only in people's heads, where it quietly leaves whenever they do.

The third, and in some ways the most interesting, is hearing the whole market. Objections, competitor mentions and feature requests come up across dozens of calls a month, but each person only ever hears their own slice of that, and a manager sitting in on a call hears one call. When the transcripts are collected and read consistently, patterns start to surface: the same competitor appearing in different territories, the same technical objection blocking deals at the same stage. This is exactly the raw material that product and engineering teams keep asking sales for, and it has historically been almost impossible to supply in any systematic way.

There is a common thread running through all three. None of them tells your reps what to do, or claims to know your market better than they do. They capture reality and make it visible, which is a humbler ambition than prediction, but a more defensible one. The judgement stays with your people. The record-keeping and the pattern-finding go to the machine, which is genuinely what it is good at.

It is also, usefully, the cheapest place to start. The transcripts, threads and notes your team already produces are the same raw material the more ambitious systems are built on: forecasting, win-rate analysis over time, the kind of knowledge that spans teams rather than stopping at departmental boundaries. Collecting it well once means that everything you want to build later stands on something solid.

That's the approach we take on every engagement, and it's why most of our systems are in daily use within weeks of the first conversation. If you'd like to talk about what this looks like in your own sales team, get in touch.

Author: Michael Wells