The starting point
A sales operations person checked signups by hand, often the next day, looking up domains and matching users to accounts. Duplicate company records were common. Existing customers were treated as new prospects, and good accounts waited a day for an owner.
What we built
Account matching on signup
Company, domain and workspace data decide whether a signup is a new company or joins an existing account. Uncertain matches wait for review.
A research brief for each company
An AI agent gathers public details on each new company, such as size and likely use, and writes a short brief on the record from sources the engineers approve.
Routing rules
The revenue lead agreed the rules that decide whether a signup goes to sales, support or customer success.
Monitoring and handover
The client’s team received the rules, test cases, monitoring and a guide to reviewing uncertain matches.
The result
Reps now reply while the customer is still in the product, with the research already on the record. Duplicate company records are caught at signup, and the client’s team runs the system itself.
- The right owner, first timeEach signup reaches the account owner or the right rep, with research attached.
- Replies while they try itReps reach new accounts while the customer is still using the product.
- One record per companyDuplicates are caught at signup instead of being cleaned up later.
Where people stay involved
The client’s operations team reviews uncertain matches and watches routing. Only confirmed duplicates are merged, with a log and a recovery procedure. Research briefs name their sources.

