Tax strategist
Relationship manager and contact data steward
Contact-Ops
In development
A relationship manager that earns its autonomy: it merges the duplicates it can prove, asks about the ones it can't, and demotes itself on drift.
Access
We run our own company's contact base on it every day, and a customer instance is built and staged. Public onboarding isn't open yet; follow the build log and you'll see the door open.
The work
What it does
- Merges duplicate contacts on its own when an exact email or phone identifier backs the match, and routes name-only lookalikes to your review queue instead of guessing.
- Runs a 5-tier trust ladder: agents earn auto-merge rights from verified outcomes, and a nightly calibration loop demotes any agent whose accuracy drifts.
- Keeps a canonical, governed registry the whole suite reads: propose-then-approve inbox, hash-chained audit ledger, 90-day reversible unmerge, and 150+ MCP tools for other apps and agents.
Watch · 1:41
Watch the Contact-Ops walkthrough
See Contact-Ops handle the work in a short, practical walkthrough.
The workflow
How it works
- 01
Import from CSV, CardDAV, or connectors. Everything lands as a proposal, nothing overwrites your data on arrival.
- 02
Agents match across email, phone, name phonetics, DOB, address, and company, auto-merging only what they can prove and queueing the rest for you.
- 03
Watch the fleet from the control plane: live roster, per-agent trust tier, pause and promote controls, and a workspace kill-switch.
Source record
A claim with a source
328 identifier-backed duplicates auto-merged with zero failures, at 99.7% measured posterior accuracy.
Keep exploring
Related employees
For humans and for agents
Bring your own agent.
Unicorn Commander is built for people and agents. Start with the connection details and the permissions for the product you want to use.
Machine discovery
/.well-known/agent-card.json
The card describes the available identity and capabilities. It does not grant anonymous access.
A useful next step
Need a guided fit?
Start with the job, the product state, and the way you want to run it.