Briefings for the people who will actually use the system.
What to do, what to escalate, and what a model or an everyday system is not allowed to do.
A briefing, not a motivational session.
Training sits under Context. It is for staff, operators and executives who need a clear account of how the systems they use fail — including language models and agents with real permissions — and what they are expected to do about it.
We do not sell a library of generic modules. The session is scoped to the environment you run: the applications, the vendors and, where relevant, the models. If the problem is technical rather than human, we will say a review is the better first step.
Sessions can be remote. The same terms apply for clients internationally. Client material used as an example is never reused to train or tune a model.
Staff
Phishing, credentials, file sharing and how an instruction hidden in email or a document differs from ordinary content. What to escalate, and to whom.
Operators
Access, logging, backups and the tools an agent can call. The difference between a scanner finding and a control that is actually in place.
Executives and counsel
What a report can and cannot support. Residual risk, authorization, and when an outside review is the right next step rather than another policy.
Scoped to your systems.
An introductory meeting sets the audience, the duration and what must not be discussed in the room. We prepare from architecture and, where authorized, from a prior audit. There is no requirement to buy a platform or a monthly subscription.
Related: Context for data and retrieval practices; AI assurance for a review of the running system.
What you receive
- A session scoped to the named audience
- Written notes the sponsor can circulate
- Escalation points that match your structure
- A clear statement of what the briefing did not cover
Request a written proposal.
Describe the system, vendor or AI deployment under consideration. We respond with the proposed scope, timeline and fee.