Silver AI made 7 mistakes. The claim team made 17.
Over a year ago, we ran a validation study with one of our customers for 30 days. It ran side by side with their claim team to prove that faster and cheaper doesn't mean less accurate.
Book a Demo
Each dot is one of the 4,876 claims Silver recommended for approval.
- 4,834 Silver and the team agreed
- 18 TPA-specific policy
- 17 claim team mistakes
- 7 Silver mistakes
All 42 differences were 
checked with the TPA
Silver recommended approving 4,876 Healthcare FSA claims and the team approved 4,834, and the 42 between them show who made fewer mistakes.
| Claims | What happened |
|---|---|
18 TPA-specific policy | The TPA's own policy denied these, but they didn't match the rest of the industry. This wasn't a problem, Silver was able to make TPA specific adjustments. |
17 Claim team mistakes 0.35% of 4,876 | The team denied these in error. The TPA got the list for training. |
7 Silver mistakes 0.14% of 4,876 | These should not have been approved. Silver corrected its logic. |
The TPA let Silver approve 
claims on its own
The TPA went live on Silver's accuracy, with each expense checked against a database of eligible products and services.
99.86%
4,869 of 4,876 recommended approvals correct
500K+
Eligible products
150K+
Eligible medical services
A free pilot measures 
Silver against your team
It uses your claims, and Silver approves nothing on its own before you see the result.
- 1
Connection to your TPA platform
- 2
30 days of your team's decisions
- 3
Each difference examined with you
- 4
Your decision on automation
Measurement 
continues after the pilot
Silver takes responsibility for the accuracy of the decisions it makes on your behalf, and a wrong approval is logged as a defect to fix.
Re-running past claims on each new version of Silver
Tracking each new version against your team
Your team's claim notes, sent hourly to Silver's engineers
grade
future of
processing
