Questions, answered plainly.

Yes. The AI reliability quiz is six questions and takes about 2 minutes. You get a reliability score, your single biggest risk, and what to do about it, on screen, with no email required. It is the fastest way to know whether a conversation is worth your time.

Take the 2-minute quiz →

Most engagements begin with the AI Reliability Audit, because you cannot repair what you have not understood. But the entry point depends on your situation: if you already know your content is the problem, we can begin with the foundation; if the foundation is sound and adoption is the gap, we can begin there. If what you need first is an AI use policy, that is available on its own.

The audit runs three to four weeks. The standalone AI use policy takes two to three. The foundation build and the enablement program are scoped to your situation once the audit has shown what actually needs doing, so their length depends on what it finds.

Almost never. The work assumes the tools you already have — Copilot, ChatGPT Enterprise, your existing knowledge platforms. When AI underperforms, the cause is usually the content it reads or the way people work with it, not the software. Fixing those rarely requires a new purchase.

No. I work on the content and the adoption: the information architecture, the ownership, the content lifecycle, and the human habits around it. I am not an ML engineer, and when a problem belongs with one, I will tell you.

Then you learn that, with evidence — which is itself worth having. The audit's job is to locate the problem, not to justify further phases. If the findings point somewhere other than my work, the roadmap says so.

Pricing is discussed in the first conversation, once the scope is clear. Each engagement is scoped to your situation, so a number quoted before understanding it would not be an honest one.

Access to the content the AI reads, and time with a handful of the people who use it. Every engagement starts with structured discovery, so the findings rest on evidence rather than assumptions. Beyond that, the burden on your team is kept deliberately light.

You do not need to prepare anything, but a short intake helps us go deeper in the thirty minutes. It asks how your AI tools are set up and where they fall short; your answers make the first conversation twice as useful.

Fill out the intake →

Yes. I am based in Vancouver and work with distributed teams across time zones, in English, Russian, Spanish, and German. On-site time can be arranged where the work benefits from it.

The work is built to outlast me. Ownership, standards, and lifecycle are placed with your people; the enablement phase leaves behind champions and materials your team can run without me. A system that depends on its consultant is not a system.