The record, its gaps, and one page for the people who asked.
Two to three weeks for one agent deployment. You provide the logs and two interviews.
timeline every relevant action, in order, with its source action record expected · authorised by · outcome · mismatch, per action gap list where no expectation, no authority, or no outcome was recorded oversight map where a human could have overridden, and whether one did mapping each finding against the logging and human-oversight duties you are asked about board page one page, plain language, for the people who asked the question
Week one
Scoping call. Log export under the data terms on the trust page. First interview: who the agent acts for, and who was supposed to be able to stop it.
Week two
Reconstruction. Code rebuilds the timeline and the action record from the logs. Where the record is ambiguous a person decides, and the pack marks each of those decisions.
Week three
Second interview to check the reconstruction against what people remember. Delivery. A walk-through with whoever has to present it.
What this is not
The review reports what the record shows and where it is silent. It is not legal advice or a security assessment, and it makes no judgment about materiality. Those decisions stay with your counsel and your board.
After the review
Most teams find the record should have been written as the agent ran. The same record keeping can run beside your agents from then on, so the next question already has its answer. Ask about it at the walk through.
Email a few lines about the deployment and what was asked of it. A person replies within one business day.