For self-funded employers, TPAs, and health plans
Bill review your members can see.
BetterBill checks hospital bills line by line and shows the evidence behind every flag. Use it before you pay a claim, after you pay, or give it to members as a self-serve tool.
Cedars-Sinai · claim review
Billed · synthetic
Evidence
Billed $6,900.00. The hospital's own price file lists a cash price of $5,143.15.
Event log
Sample review. Synthetic data. Pick a flag.
Measured on synthetic bills, Oct 18, 2026
0 of 22
seeded errors found
Planted by the team across 11 bills.
0
false flags on clean bills
Checked on 3 clean bills.
0.000 ms
median audit time
A deterministic rules engine. No model in the loop.
0 of 3
published findings matched
In-scope reporter findings from 5 recreated cases.
One review, three places to use it.
Check itemized hospital bills before a claim is paid. Flags come with the line, the rule or price, and the source.
Look back at paid claims to find duplicates, unit errors, and prices above the hospital's own published rates.
Members snap their bill and see what looks wrong. They approve every letter and call, so nothing happens behind their back.
How a review runs.
Bill in
Photo, PDF, or text. Every line read and confirmed.
Audit
Duplicates, units, insurer statement, published price.
Evidence
Each flag carries its tier, source, and capture date.
Event log
Append-only. Every approval is a recorded event.
Evidence you can audit.
- Confirmed from your bill
Arithmetic on the submitted documents
- Billing rule
A cited billing rule with its source and date
- Hospital's published price
The hospital's published price file, with capture date
- Estimate
A labeled estimate, never counted as confirmed
Every action is an event
Example event. Synthetic data.
- at
- 2026-10-17T23:10:00Z
- actor
- user
- type
- action.approve
- summary
- Approved dispute letter
Events are append-only. Corrections are new entries, so you can see who changed what and when.
Measured, not promised.
Seeded errors found, by check
- Duplicate charges4 of 4
- Unit counts5 of 5
- Bill vs insurer statement2 of 2
- Published price10 of 10
- Missing itemization1 of 1
The team planted these errors on synthetic bills. This shows the rules work as designed. It is not a claim about real-world accuracy.
Against published cases
1 of 3
in-scope reporter findings matched, across 5 recreated Bill of the Month cases.
We also raised 3 extra flags. We report the misses with the hits, because a result you can't check isn't one.
See where flags cluster.
A plan view ranks hospitals by how often reviews raise flags, so you know where to look first. This map is a mock-up.
Illustrative. Made-up data.
Selected
Hospital C
Southern California
52flags per 100 bills
Hospitals A to H and every rate are made up to show the view. They are not measurements. A flag is a question to ask, not a finding of error.
Built to plug in.
Typed API
Every step is an API call: create a case, read a bill, run the audit, draft a letter. An OpenAPI description is available.
Case events
Each case keeps an append-only event log you can read back. Webhooks are on the roadmap.
Intake today
Photos, PDFs, and pasted text. Claim file intake is on the roadmap, not built.
Security, stated plainly.
- Synthetic data only
- Model and phone keys stay on the server
- Every letter or call needs a recorded member approval
- Calls identify themselves as an automated assistant
- No HIPAA program or business associate agreements
- No SOC 2 report
- No real member data until those are in place
We'd rather tell you what's missing than show a badge we haven't earned.
Bulk plan.
One monthly plan for insurers and plans. This is demo pricing in the Polar sandbox, so no real money moves.
a month (demo pricing)
Pilot partners wanted.
We're looking for one or two self-funded employers, TPAs, or plans to shape a pilot on synthetic or de-identified data.
A pilot includes
- A review of sample bills you choose
- Evidence reports for every flag
- A member flow your team can try