Know what an agent can do before you use it.

Why it matters

You choose an agent on evidence, not on its advertising. It is tested on the kind of work you need, and someone else can repeat the test.

An example

An illustration of how it is designed to work.

1. A buyer needs a contract translated, and a translation agent claims that skill. The assessment names the agent and the assessor.
2. The assessor examines the risk families chosen for that kind of translation. Every family examined is recorded, so finding no problem is different from not looking.
3. For each problem found, the assessor records the family, the exact conditions met and the evidence behind the finding. Any proposed remedy must point back to those conditions or that evidence.
4. The complete result is put into one fixed form with the assessor's identity. A fingerprint of that record makes later changes visible.
5. The result is marked self-assessed unless the test is expressly identified as independent. Different names alone do not prove independence.
6. Anyone who runs the same inputs gets the same record and fingerprint, so the result can be reproduced and checked.

The buyer gets a traceable result for the work at hand, including who tested the agent and what evidence supports each finding.