AI-Generated Audio Summary

We have been building AI tutors for years, at least four. Not as demos, and not as novelties, but as working systems that help real people learn difficult things. Verónica, the tutor many of you already know from our site, is one of them. Everything those years taught us, how to ground a tutor in real content, how to keep it honest, where it helps and where it should stay quiet, now lives inside a platform we call Praxitude AI.

This post is about one of the first courses we built on it, and about something the expert we partnered with asked us to make unmistakable. After he saw how capable the tutor had become, he asked us to be clear about one thing: this is a tool for studying, not for doing the job. That distinction is the part worth explaining, but first it helps to know what Praxitude AI is actually for.

What Praxitude AI Is Built to Do

We built Praxitude AI to create scenario-based learning, the kind where a learner does not simply recall facts but works through realistic situations. In its fullest form, the platform measures two things at once: the learner’s language and their knowledge of a domain, both the theory and how they apply it. That dual measurement, language and competency together, is the heart of what Praxitude AI was made for. You can read the full picture on the Praxitude site.

But we built it to be modular and adaptable. Real training needs do not all look the same, so the platform flexes. It can power a full scenario-based course that assesses language and applied competency, or a focused certification-prep course, or a professional-development program, and it shapes itself to the requirement rather than forcing every subject into one mold. That adaptability is exactly why we can partner with experts across very different fields.

One thing stays constant in every course we build, whatever its shape. Praxi, our AI tutor, is embedded everywhere inside it. Praxi is the through-line, the support that sits beside the learner at every step, explaining, connecting, and answering, always grounded in the expert’s approved material. Praxi grew directly out of everything Verónica taught us. That constant, an expert-grounded tutor available at every moment, is what this story turns on.

How We Work With Experts

The model behind Praxitude AI is a partnership. We bring the expertise in building AI learning systems. The expert brings the authority in their field, the years of teaching a subject and knowing exactly where learners struggle. They own the content and the judgment. We turn it into a course that teaches at scale without losing accuracy, with Praxi available at every step.

We are deliberate about this because we are not the authority on every subject, and we do not pretend to be. Our job is to give people who are the authority a way to teach far more learners than they ever could one classroom at a time, and to do it responsibly. This course is one of the first times we ran that model end to end with an outside expert.

The Challenge Guillermo Brought Us

Our partner is Guillermo Ventura, a veteran in PLD/FT compliance in Mexico, the field of preventing money laundering and the financing of terrorism. He prepares professionals for a national certification, and he knows precisely where candidates get stuck.

The difficulty was not motivation. It was volume. The material behind this certification is enormous: laws, regulations, official guidance, real cases, hundreds of pages of dense legal PDFs, and the reasoning that connects one rule to another. The problem was never that people would not read. It was that the sheer mass of the source material makes it hard to hold the whole picture in your head. Guillermo wanted a way to make that training more approachable, so we sat down to design one together.

What We Built

We organized the full course first, under his guidance and following the structure he laid out. He is the author and the authority; the sequence, the emphasis, and the corrections were his. Then we did what Praxitude AI does in every course: we put Praxi at the center of the learning experience, grounded entirely in his approved material.

Under the hood, that tutor runs on what is called retrieval augmented generation, or RAG. In plain terms, Praxi does not answer from memory or from the open internet. When a learner asks a question, it retrieves the exact passages from the approved content and answers only from those, pointing back to where the concept lives in full. Around that we built heavy guardrails and rich, context grounded prompting, so the tutor stays inside the lines: it does not invent, it does not drift into opinion, and when the source does not cover something, it says so.

The design principle matters here. The system still invites you to read the source. It does not replace the reading; it opens a conversation with it. A candidate can sit with the actual regulation and, when a passage is confusing, ask Praxi to explain it, connect it to another rule, or show where a term is defined.

What Surprised Us Both

When Guillermo and I finished building the whole thing together, we were both genuinely surprised by what the system could do. Guillermo is the authority in this field, and Praxi still showed him things. It surfaced connections across different laws and documents that are easy to miss, and it read the fine print carefully enough to point back to the specific provision behind an answer. He knew this material cold, and he was struck by what the tutor could see across the entire body of content at once.

That is the part of this work that still moves us. When you embed a capable, well-grounded tutor in an expert’s course, the way we first learned to do with Verónica and now do with Praxi, the results can be genuinely remarkable.

The Line Between Learning and Doing

And that is exactly why Guillermo suggested a disclaimer.

His point was simple. Because Praxi knows this body of law so thoroughly, you can pose a hypothetical to it and it will reason through a response, grounded in the regulation. For someone studying for the certification, that is enormously useful, a way to test your understanding against the material and see how the rules apply. But responding to a hypothetical in a course is not the same as doing the actual work that a certified compliance professional does, and Praxi was never built to be that.

So we said so plainly, right in the course. Praxi is here for educational purposes, to help you learn, practice, and prepare for the exam, and not as a substitute for a professional’s judgment, current regulation, or human review.

This part is easy to misread, so it is worth being precise. We did not limit what Praxi can do. We were clear about what it is for. The disclaimer is not us dialing the AI down; it is us stating, in plain sight, that this is a learning tool, not a tool designed to do the certified work of an expert in the field. That distinction, between what a system is capable of and what it is meant for, is the part that earns trust. Deciding it together with the expert is the half of this work that matters most.

Where This Goes

This is the Praxitude AI model, and it is why we can work with a compliance specialist in Mexico, and next with an expert in another country and another field, without pretending we are the authority on every subject. The platform adapts to the training need, Praxi supports the learner throughout, and the expert stays in control. The technology scales their reach; it never overrides their expertise.

We are opening this partnership to more experts who want to turn hard-won knowledge into a course that reaches more learners, without becoming a software company to do it. If that is you, this is the kind of partnership we are built for. And if we ever build you something that looks a little too capable, do not be surprised if we are the ones who suggest a disclaimer. That is not us being modest about the technology. It is us being serious about the responsibility.

Let’s Talk

Are you an expert with hard-won knowledge and no interest in becoming a software company? That is exactly the kind of partner we build for. Let’s talk about building it in Praxitude AI.

Praxi and Verónica are AI tutors built by Enabling Learning. They are designed to support learning, practice, and preparation. They are not a substitute for professional judgment, current regulation, or certified expertise.