Salespeople spend a third of their time on this
Reading the website, checking recent news, looking at the CRM history, working out who else in the business has spoken to them. Fifteen to twenty-five minutes a call, and it is essential work performed badly under time pressure.
It is also entirely automatable, because it is gathering and summarising rather than judging.
What the brief contains
- Company summary — what they do, size, sector, in three lines
- Recent signals — funding, hiring, product launches, news
- Relationship history — every previous contact, from the CRM and inbox
- Likely fit against your typical customer profile
- Three suggested questions, specific to them
Delivered to the CRM record an hour before the meeting, and to the salesperson's inbox. Nobody has to remember to request it.
Why it is low risk
Nothing goes to the customer. The salesperson reads it, forms their own view, and any error is caught in the first minute of the conversation.
That makes it an unusually good first AI project: real value, immediate feedback, no exposure.
Sourcing matters
- Public sources only, cited so claims are checkable
- Your own CRM and inbox, which is the richest source and the most neglected
- No inference presented as fact — “appears to” where the source is thin
- A clear note when little was found, rather than padding
The measurable effect
Time per call preparation, calls per rep per week, and how often the brief is opened. The last one is the honest measure of whether it is useful.
In practice open rates above 80% within a month mean it has landed; below 40% means it is not saying anything the rep did not know.