Compliance AI Operations
Evaluate the recordability agent's accuracy and generate the proactive compliance digest.
Recordability Evaluation
Runs the recordability agent against a labeled golden set and scores it.
Run evaluation
—
0/12 cases passed
llama-3.3-70b-versatile
1 case failed — gc-01: the agent called a bandage recordable.
gc-01
First aid only (bandage).
exp: not / got: not
gc-02
Sutures are medical treatment beyond fir…
exp: recordable / got: recordable
gc-03
Prescription medication is beyond first aid.
exp: recordable / got: recordable
gc-04
No treatment beyond first aid; no days away.
exp: not / got: not
gc-05
Fracture + days away from work.
exp: recordable / got: recordable
gc-06
Splinter removal by simple means is first aid.
exp: not / got: not
gc-07
Not work-related.
exp: not / got: not
gc-08
Prescription medication + restricted work.
exp: recordable / got: recordable
gc-09
Work-related hearing loss case.
exp: recordable / got: recordable
gc-10
Loss of consciousness after chemical exposure.
exp: recordable / got: recordable
gc-11
OTC medication at nonprescription strength.
exp: not / got: not
gc-12
Needlestick from a contaminated sharp.
exp: recordable / got: recordable
Compliance Digest
Generates a prioritized digest from live signals and delivers it to your inbox.
Generate digest
No digest generated yet.
Overdue actions: 3 corrective actions are past due
Facilities at risk: 1 facility (Unassigned) needs attention
DART vs target: Current DART rate (0) is below industry benchmark (1.5)
Delivered to your notifications.