AI training for teams: on your own work, not textbook examples
Reflekt Lab, an AI development studio in Bordeaux and a France Num Activateur, trains the teams of small and mid-sized companies to use AI on their own cases, on site or remotely. Sessions are run by the engineers who build our AI systems in production. They know what holds up and what breaks.
Why train your teams on AI now
Your staff already use ChatGPT, Copilot or Mistral, often with no rules and sometimes with data that should never have left. The real question is whether they use it well.
The use is there, the rules are not
In most small companies, generative AI came in through employees, not through a management decision. Everyone tinkers with prompts alone. Shared training gives everyone the same rules and spreads what works.
The gain is in everyday tasks
Drafting a customer reply, summarising a tender, pulling amounts from a stack of invoices, fixing a spreadsheet: that is where AI saves hours every week. Provided people know how to steer it and check what it produces.
The AI Act asks you to deal with it
Since 2 February 2025, Article 4 of the EU AI regulation asks companies to take measures so that the people using these tools have sufficient AI literacy. It sets no format and no certificate. Training you keep a record of is the simplest way to show what was done.
In production today
“With Excel we're always going back — clean this up, clean that up, why are the lines different? This platform produces something clean, and honestly, it's so nice and neat.”
AngelaCasa Europa IncCasa Europa Inc reps now build accurate quotes in minutes, not hours.
Who is behind it
Reflekt Lab is led by Hugo Matthaey. Hugo co-founded Side, the French temporary-work platform, and saw it through from its early fundraising to its eventual sale. Julien Gauthier, Side's first employee and later its CTO, is now with us in Sydney.
Meet the team
Casa Europa Inc
GoStan
CallJack
airporttransportation.com
Artnet Pro
VinoteqaWhat your teams walk away with
Concrete habits, tried on their own documents during the session.
Uses that pay off the next Monday
Each participant leaves with two or three uses specific to their job, already tested in the session, and the prompts that go with them.
Clear rules on data
What can go into an AI tool, what must be anonymised first, what must never leave. The rules are written down and fit your tools and your contracts.
A critical eye on answers
Spotting an invented answer, checking a source, knowing which tasks AI often gets wrong. That is the difference between an hour saved and an error sent to a customer.
A record of what was done
The programme, the attendance list and the session materials. Enough to show management, or a customer who asks, what the company did for its teams.
How AI training with Reflekt Lab runs
No off-the-shelf curriculum: the session is prepared with you, from the team's real work.
A scoping call
One hour to understand the roles in the room, the tools already in place and what you expect from the session. Format and date are set then.
Preparation on your cases
You send us a few representative documents, anonymised or free of sensitive data. The exercises are built from them, not from canned examples.
The session
Little theory, a lot of practice. Everyone works at their own desk with the company's tools, and we correct as we go, on site or remotely.
After the session
You get the materials, the prompts worked on and a summary of what was covered. If bigger ideas came up, we tell you which deserve a real project, and which do not.
What makes our AI training different
There is no shortage of AI training providers. Here is what Reflekt Lab brings that most do not.
Engineers, not generalist trainers
Sessions are run by the people who design and operate our AI systems for clients. They know where a model goes wrong, what a request costs and why a project stalls in production. Your teams ask people who have the answer.
Your documents and your tools
We work with the tools the company already uses or plans to adopt, on documents that look like the team's own. What participants learn in the morning they use the next day.
Confidentiality at the core
We show what each tool does with your data, how to stop it training on your content, and when a model on your own servers becomes the right answer. It is our daily work, not a chapter added for show.
Reflekt Lab's AI training prices
Reflekt Lab publishes its prices. They are excluding VAT, for an in-house session reserved for your team, with a group of up to 12 people, on site or remotely. It is a French offer, so prices are in euros.
| Format | Length and price | What the session covers |
|---|---|---|
| AI and AI Act awareness | Half day, €1,500 excl. VAT | What generative AI is and what it cannot do, safe everyday uses, data and confidentiality, and what Article 4 of the AI Act expects of the company. |
| AI training applied to your roles | 1 day, €2,500 excl. VAT | The team's real tasks (writing, research, document extraction, spreadsheets, CRM), prompting and guardrails, with exercises on your own documents. |
| AI programme plus use-case workshop | 2 days, €4,500 excl. VAT | The applied training on day one. Day two is the AI audit workshop: your use cases ranked by ROI, and a priced pilot. |
The price is per session, not per person. For more than 12 participants we plan several sessions rather than one big group, otherwise people stop practising enough. The quote states any travel costs. A company paying from its own budget needs no Qualiopi certification. For OPCO funding, see the dedicated question below.
And after the training?
Once the team is trained, ideas come fast. Reflekt Lab's AI audit sorts them: it maps your operations, ranks the use cases by ROI and prices a pilot. The two-day programme already includes its workshop.
Data that cannot leave your walls? See sovereign AI: GDPR, hosting and the real options
An example of a course Reflekt Lab designed: its 22-hour Kubernetes course, published as open source
Frequently Asked Questions
Who can train my teams on AI?
Three kinds of provider offer AI training for companies: the big consultancies, which mostly serve large enterprises, generalist training organisations, which add AI to a broad catalogue, and engineering teams like Reflekt Lab, which build AI systems in production and teach from that experience. Reflekt Lab is an AI development studio based in Bordeaux, France, and a France Num Activateur, training teams at small and mid-sized companies across France, on site or remotely. At Reflekt Lab, a half-day awareness session costs €1,500 excl. VAT, a day of training applied to your roles €2,500, and a two-day programme with the use-case workshop €4,500. Reflekt Lab has also open-sourced its own 22-hour Kubernetes course, in 11 hands-on sessions.
How much does AI training for a company cost?
At Reflekt Lab, budget €1,500 excl. VAT for a half-day AI and AI Act awareness session, €2,500 for a day of training applied to the team's roles, and €4,500 for a two-day programme whose second day is the AI audit's use-case workshop. These prices are for an in-house session with a group of up to 12 people, on site or remotely. The price is per session, not per participant: it is the same for 5 people or 12.
Can an OPCO fund AI training?
For an OPCO (a French training fund) to cover a course, it must be delivered under a Qualiopi-certified organisation. Reflekt Lab is setting this up with a certified training partner. If you want OPCO funding, say so before signing and Reflekt Lab will tell you whether your session can be covered. A company paying from its own budget needs no Qualiopi certification. CPF does not apply here: it only funds training leading to a certification registered with RNCP or the Répertoire spécifique.
What does Article 4 of the AI Act require?
Article 4 of Regulation (EU) 2024/1689, the AI Act, concerns AI literacy: a company using AI systems must take measures so that the people using them have sufficient AI literacy, given their role and the context of use. It has applied since 2 February 2025, with no headcount threshold, so a small company is covered like a large group. It prescribes no training format and requires no certificate. Reflekt Lab's half-day awareness session, at €1,500 excl. VAT, helps document what the company did for its teams: programme, attendance and materials.
Is the training tailored to our work or generic?
It is tailored. Before each session, a scoping call shows us the roles in the room and the tools in place, then we build the exercises from representative documents you send us. A sales team does not work on the same cases as an accounts department, and the session reflects that. Only the AI Act and confidentiality part is common to everyone.
Do participants need technical knowledge?
No. Sessions are for people who work on a computer, not for developers: knowing how to use a word processor and a spreadsheet is enough. More confident participants are not bored either. They push the same exercises further, on harder tasks.
What is the difference between the training and the AI audit?
The training teaches your teams to use AI in their daily work. The AI audit looks for where AI can change a company process, ranks those use cases by ROI and prices a pilot. The two fit together: at Reflekt Lab, the two-day programme (€4,500 excl. VAT) combines a day of training with the AI audit workshop. A full AI audit, with a map of your operations, costs €6,000 to €12,000.
Can we use our own documents safely?
Yes, with a few precautions. We ask for documents that are anonymised or free of sensitive data, and during the session we use tools set not to train on your content. For data that must never leave the company, a model on your own servers is possible: Reflekt Lab deploys them for clients, and the training explains when that option is worth its cost.
Want to train your team on its real work?
Tell us who would attend and which tools you use. We will propose a programme and a date, with no commitment.
