Benchmarks

PACT

Benchmark · 1 item since Sep 2026

Whether enterprise assistants apply compliance rules correctly across twelve regulated domains under user pressure.

Items

  1. Sep 18

    PACT tests enterprise assistants under user pressure

    A pressure-tested compliance benchmark across twelve regulated enterprise domains finds even the strongest assistants misapply a governing rule on 6-10% of items, and ordinary user pressure raises the rule-violation rate by 65% on average.