Loading...
Menu
Loading...
Executive literacy through to hands-on engineering, taught on your own documents and processes. People leave having solved something real, not having watched a demo.
The most common way corporate AI training fails is by aiming a single session at everyone. The executives find it too operational, the engineers find it shallow, and the people who actually do the work get examples about a fictional company. Everyone leaves politely and nothing changes.
So we separate them. Leadership needs enough grounding to judge an investment, ask a vendor a hard question, and understand where the regulatory and reputational exposure sits. That is a different session from teaching an operations team to restructure how they handle a daily queue.
Engineers need something else again: retrieval and chunking, evaluation, why agents fail in production, and where the model is the wrong tool. That audience notices immediately when a trainer is reading slides, so those sessions are run by people who build these systems for a living.
Programmes
Half a day for leadership: what the technology can and cannot do, how to read a vendor claim, where the risk actually sits, and how to judge an internal proposal.
One to two days for the people doing the work, rebuilding their own recurring tasks with AI in the room. They leave with something already working.
Depth for engineers on retrieval, evaluation, agent design and failure modes, spread over weeks so the material gets applied between sessions.
Learn more →Training alone rarely changes behaviour. Where a team is genuinely changing how it works, we stay involved through the transition.
Learn more →We start from your material. Before a session we look at the documents, processes and systems the group actually works with, and the exercises get built on those. This is the part that decides whether anything survives the week.
Sessions are mostly hands-on. People spend the time working on their own tasks with support in the room, rather than watching somebody else's screen.
We agree in advance what participants should be able to do afterwards, and we check against that. A satisfaction score measures whether the day was pleasant, which is not the thing you are buying.
Three distinct audiences, and mixing them is the usual mistake. Executives need enough literacy to make investment and risk decisions. Operational teams need to use the tools well on their own work. Engineers need depth on retrieval, evaluation and agent design. A single session aimed at everyone lands with nobody.
Exercises are built on your own material: your documents, your processes, your systems. Generic training produces people who can prompt a chatbot about a fictional company. Training on your own data produces people who solved a real problem during the session and can repeat it on Monday.
An executive session is typically half a day. Team workshops run one to two days. Technical programmes for engineers run longer and are usually spread across several weeks so people can apply the material between sessions.
Both. Executive sessions and technical workshops work well remotely. Where a team is changing how it works day to day, being in the room is usually worth the travel, and we run those on site.
French and English. For teams that work across both, we mix, and materials can be supplied in either.
We agree beforehand what people should be able to do afterwards, and check against that rather than against a satisfaction score. Enjoying a session and changing how you work on Monday are different outcomes, and only the second one is worth paying for.
The programme is built around the audience and the work they do. A short call is enough to shape it and say what it would take.
Discuss a programme