Services / Clio data hygiene
Clean up duplicate Clio contacts without losing a thing.
Years of intake leave every firm's Clio full of duplicate people: same client, two records, notes and documents split between them. We built a system that finds them, asks a human to confirm each one, and merges with a paper trail that makes every merge reversible.
The problem
Duplicates aren't cosmetic. When one client exists twice, their history is split in two, and whichever record a staff member opens is missing half the story. Phone matching, conflict checks, and reporting all quietly degrade.
The reason nobody fixes it is that merging is scary. There's no undo. Combine the wrong two people, or lose a field along the way, and you've damaged a legal record.
How we build it
Detect on a schedule
A weekly scan looks for same-person signals across the firm's contacts and queues candidate pairs. Detection never writes anything. It only looks.
Three independent checks before anything moves
Identity: the signals confirm it's the same human. Survivor: the record with the most case work is the one that remains. Freshness: fields are combined, and a legal name, date of birth, or SSN is never silently overwritten. All three are re-verified at the moment of merge, not just at detection time.
A human clicks every merge
A quarterly review email lays out each duplicate as a card. Nothing merges without an explicit confirmation click on a signed, expiring link. The AI in the pipeline has one power only: setting a pair aside as different people, like spouses sharing a phone. It cannot cause or approve a merge.
Reversible by design
Because Clio has no merge API, the merge reassigns everything off the duplicate first and deletes only when nothing references it. Before deletion, the full record is snapshotted into an append-only reversal manifest, with Clio's own recovery bin behind that.
Where it runs today
Built for a Social Security disability firm as the only system permitted to write to their Clio contacts. It shipped with a global kill switch defaulted to off, and no merge ran until the first one had been reviewed end to end.
Common questions
What if it tries to merge two people who aren't the same?
Three things stand in the way: the identity checks, the AI triage that flags family members who share contact details, and the fact that a human has to confirm every single merge. And if a wrong merge somehow got through all of that, the reversal manifest holds everything needed to put it back.
How does it handle spouses or a parent and child sharing a phone number?
That's the exact case the triage step exists for. Shared phone or address with different identities gets set aside as “different people” and never shows up in the review again.
How much of our time does this take?
One review email a quarter. Each duplicate is a card with the evidence laid out, and each decision is one click. The finding, checking, merging, and record-keeping is the system's job.
How many of your contacts are the same person?
Book a free audit and we’ll walk your setup and tell you what this would look like at your firm, including what it would save.