Census intake, re-engineered

Your clients can’t send you a bad census anymore.

Every census is validated the moment it’s opened — in the browser, before it ever reaches your queue. Errors are caught at the desk that can actually answer for them, fixed where fixing is cheap, and the file that lands with your team is already proven clean.

Runs entirely in the browser · participant data is never uploaded · deterministic rules, not AI

Incoming census filesFive plans, one morning
↓ Validation gate ↓
Your teamIntake queue
Nothing invalid gets through. Corrections happen where the answers are — most never need a human email at all.
Where corrections actually happen

Three desks. The work goes to the right one.

The program fixes what it can prove

A stripped leading zero, a mangled date format, a percentage entered as a decimal — mechanical damage gets a one-click, audited repair. No email, no waiting.

Your administrator gets an alert, not a hunt

Everything else is flagged with the row, the field, the value, and a corrected-value box on every single card. What used to be an afternoon of hunting is a two-minute review.

The sponsor answers only what only they know

Questions no file can answer — whose deferral is right, which birth date is real — go into one context-rich packet. One email out, one answer back.

The line item that isn’t on any invoice

Census cleanup is a labor tax. You pay it every plan, every year.

No administrator bills for it and no client sees it. A file arrives, something is off — a stripped SSN, a hire date before a birth date, compensation in the wrong column. Now it’s an email, a wait, a second file, a second review. It reads as the cost of doing business right up until you multiply it by your whole book.

Run it against your book

These are your assumptions, not our statistics — set them to what you actually see.
$90,000
absorbed every single year in staff time spent making client censuses usable — before a single compliance test is run.
1,200
Hours annually
0.7
Full-time staff
600
Files corrected
Proven clean, not just un-flagged

24 check families. 53 rules. Every record, every time.

Identity, dates, hours, compensation, contribution limits, employer-money logic, ownership, and the shape of the file itself. Every check is shown with its outcome — passed, flagged, or not applicable — so a clean file is demonstrably clean. Deterministic code with published IRS limits: the same file gets the same verdict on any machine, any day.

SSN format & validityDuplicate peopleDates in orderAge at hireRehire logicHours ranges§401(a)(17) comp cap§402(g) limitCatch-up rulesMatch logicPlan entry logicOwnership & HCETotals-row detectionPrior-year cross-check…and a one-click fix path on every card
Security you can verify yourself

The census never leaves the browser. Provably.

There is no server, no upload, no analytics, and no AI reading participant data — the whole application is one page of inspectable code. Open your browser’s network tools while validating a file: after the page loads, zero requests. That’s the entire security model, and anyone — including your CCO — can check it.

The census filechosen from the desk it lives on
Browser memoryparsed locally
Validation53 rules, in-browser
Clean exportssaved straight to that device
 

See it run against a real census.

Two minutes with a messy file makes the case better than any page can.

See it in action →
Start the email in your own mail app: