Trust Agent: free AI education, and why it is staying free

Trust Agent is Mickai's free AI education site. Right now it holds more than sixty course workbooks, and every one of them is free to download. No card, no account to create, no trial that quietly turns into a subscription later. I built it that way on purpose, and I want to explain why, because "free" gets used carelessly online and I would rather be specific about what it actually means here.

The gap I kept running into

Most AI education I came across sits at one of two extremes. There is a lot of very shallow content: "what is a chatbot" explainers that tell you AI exists and not a great deal more. And there is a lot of genuinely good material that assumes you already know how to program, or already understand what a token or a parameter is before you have started. Between those two, there is very little for someone who wants to go further than "what is a chatbot" but has no reason to learn Python first. That gap, not a business plan, is where Trust Agent's workbooks came from.

From your first prompt to frontier models

The library is built as a path, not a pile of unrelated articles. It starts with a plain-English introduction to what AI actually is and how to write a useful first prompt, and works up through how agents work, running models yourself, and keeping up with new frontier releases, without assuming you arrived already knowing the vocabulary. Each workbook is a PDF: an explanation, a worked example, an exercise you actually do, and a short set of questions at the end so you can check whether it landed. It is yours to keep, print and write in. The starter workbook is a fair example of the tone: it explains, it does not perform.

The library is organised into four levels, and the honest advice on the site is to start wherever you actually are rather than where a course assumes you should be. Someone who has never opened an AI tool and someone who wants to run a model on their own hardware are not reading the same workbook first.

Why it has to stay free

I could charge for this. There would be nothing unusual about that; plenty of good AI courses do. But charging for the material would work against the actual reason I built it. Trust Agent exists because I think AI earns trust by being understood, not by being impressive. Someone who understands what a model is actually doing, where it can be confidently wrong, and what "agent" really means once software starts taking actions on a person's behalf, is in a far better position to keep that AI under their own control. Putting that understanding behind a paywall would mean the people least able to pay are also the people with the least ability to check what AI is doing around them. That is the opposite of what I am trying to build, on this site and in Mickai's own product.

What finishing a workbook gets you

Each course ends the same way: a closing page inviting you to generate a certificate with your name on it, to add to LinkedIn or a CV if that is useful. Right now that certificate is self-declared rather than the product of a graded exam: the pass-marked check behind it is still being built. I would rather tell you that plainly than let the page imply a test that is not there yet. It is not a sales funnel dressed up as education. There is nothing to buy at the end of it.

If you want to see the same thinking applied to a working product rather than a workbook, that is what Mickai is building: software that is meant to explain what it did and why, and to stay under the user's control rather than asking them to trust it blindly. The current build is open to try at mickai.co.uk/beta.

The library launched at the end of September 2026, and the real number of workbook downloads since then, checked directly against the site's own tracking rather than estimated, is in triple figures already, with no advertising spend behind it. That is a small number in absolute terms, and I am not going to dress it up as more than it is. But it tells me the gap was real: people are finding these workbooks through search, not through a marketing push, which is the only kind of early signal worth taking seriously.

Trust Agent will keep growing. More workbooks are being written. The free part of it is not going anywhere, because it is the whole point of the format, not a promotional phase before the real product arrives.

About this account: Mickarle Wagstaff-Irons - Micky Irons, full name Mickarle Sean Junior Wagstaff-Irons, is Mickai's founder and built Trust Agent as part of Mickai's education work. Micky's biography. Unified and Trust Agent are part of the same Mickai publication family; these links are not independent endorsements.