That is what we build. Practical workshops, by role, in Quebec French. ChatGPT — or Copilot, Claude, Gemini, if that is already your tool — goes into the real work. Not into a demo that floats above it.
The problem is almost never the tool
Most companies already have something. A ChatGPT licence. A Copilot “for everyone.” A Gemini account stuck to Google. What stalls it is the fog.
Nobody decided how it gets used, on which tasks, with which limits — or how that gets passed on. Result: three people move ahead, the others watch, and six months later you wonder whether “the AI training” changed anything.
We do not train people so a box gets checked. We train so a process holds. The full method — diagnostic, training, playbooks, adoption — is on Approach. If you want an AI consultant in Quebec to frame the mandate first, start there.
What the training is not
Not a tour of “the 50 prompts of the moment.” Not a 90-minute webinar forgotten in two weeks. Not a generic “AI training for SMEs” catalogue that talks about everything and trains nobody.
People practice. On your documents. We show the move, they do it, we correct it. They leave with a routine, not only a burst of inspiration.
Leadership and the floor do not get the same workshop — on purpose. A VP does not need the same exercise as a customer-service team. A firm does not have the same guardrails as a marketing team.
By role, not by slide
ChatGPT is not used the same way in every job. We split it.
Leadership. A clear read of what is worth an investment, what can wait, and how to talk about it inside the company without selling a dream. A 2-to-3-hour briefing: what the tool actually does, what it invents, where it creates risk, where it creates time.
Operations. Summaries, follow-ups, checklists, file handoffs. How to paste the right context without pasting the whole cabinet. How to check an output before it goes out.
Finance and admin. A first pass on a report, a tracker, a meeting recap. What stays a human check. What never goes into the tool.
HR. Job posts, internal notes, interview prep — with a tight frame on personal information. ChatGPT does not score your candidates on its own. A human rereads. Always.
Sales. Objections, follow-ups, first drafts of a proposal. Your company’s tone, not the tool’s. Quality checks before it goes to the client.
Customer service. Opening a ticket, restating it, a base of replies. What can leave quickly, what needs a human, what we never invent.
Each workshop starts from your tasks. Not from a scenario invented for the demo.
In Quebec French, on this week’s work
The workshops, the playbooks, and the policies are written in Quebec French. We can work in English when a mixed team asks.
“On your files” means your email templates, your standard reports, your real objections, your tickets. We do not arrive with a fictional data set. We arrive with the week’s work — anonymized when it needs to be, framed for Law 25.
People see the move on something they recognize. That is why it holds better than a workshop that floats above the work.
Formats: sprint, program, year-round
We fit the size of the company. We do not ask the company to fit us. Private ChatGPT training for one team, or a multi-cohort path for a larger organization.
30-day sprint. A short diagnostic, one team workshop, one process written down. For an SME that wants to stop circling. One cohort, one role, one move that becomes a way of working.
90-day program. A roadmap, training by cohort, playbooks on two or three key processes, coaching the leads. The format we get asked for most. This is where ChatGPT stops being “the thing two curious people do” and becomes a team routine.
Year-round support. For organizations rolling this out across several teams: governance, a training calendar, a quarterly review, useful watch. New people get a path. Internal leads keep the wheel.
Workshop lengths, as we already deliver them: a leadership briefing (2 to 3 hours), role workshops (a half-day or a day), materials the team can reuse. Training alone almost never holds — which is why the lead coaching and the 30-minute rituals, weekly or biweekly, are described on Services.
Private, or in a group? Almost always private, at your place, on your files. A room of several companies does not anchor a process. An internal cohort does, when you have several teams.
Montreal, the regions, remote
Workshops happen in person in Montreal and in the regions when the trip is worth it, or remotely when that is more useful. ChatGPT training in Montreal, elsewhere in Quebec, or online: the test is whether people can actually practice. A room with their files often opens better than a Zoom. A well-framed Zoom beats a trip for three people.
Free ChatGPT, company ChatGPT, Copilot — we settle that before the workshop
We have no allegiance to a platform. We look at what you already have (Microsoft 365, Google, Slack), how sensitive the data is, and how comfortable the teams actually are.
Training everyone on free ChatGPT, then pasting bits of client files, is exactly the shortcut we avoid. The workshop also says which workspace, which settings, what not to paste, and who may create an internal GPT. A sober comparison: ChatGPT, Copilot, Claude, Gemini, line-of-business tools. Avoid the double (or triple) subscription. If Monday’s tool is already Copilot in Microsoft 365, see Copilot for Quebec companies.
The use frame — public, internal, confidential, forbidden — is worked in the training, not six months later. See Law 25 and AI.
The 1% and Services Québec: we tell you if it might qualify
Training catalogues like to show logos. We will not do that here.
This site does not promise that our workshops qualify for the 1% (the skills law) or for a Services Québec measure. That depends on your situation, on the shape of the activity, and on rules a web page does not settle.
What we do: on the first call, we tell you whether, from our side, it might qualify — and what you would still have to check on yours. We do not promise that it qualifies. We do not put up a logo to pretend.
What people bring back Monday morning
Not a certificate on the wall. Moves.
A first pass on an email that leaves cleaner, reread by a human. A file summary the manager no longer rebuilds from zero. A proposal that has a skeleton, instead of a blank page. A team that knows what it is not allowed to paste into ChatGPT — and still uses it, instead of boycotting the tool out of fear.
We measure simply: real use, time recovered, perceived quality, irritants. We adjust. Then we leave the internal leads to keep the wheel.
What you receive
- A workshop (or a path) built on your tasks, not on a template
- Materials the team can reuse
- At least one written process: when ChatGPT gets opened, what it is given, what gets checked, what is never delegated
- A use frame clear enough to pin next to the screen
- Coaching for the leads, when the 90-day or year-round format includes it
You will not get: an orphan pilot, training that floats above the work, or pressure to buy a tool we resell. If the mandate has no internal sponsor and no first process to anchor, we say so immediately — before we invoice.
What it costs
It depends on how many teams and how long. A scoping sprint is not a year-long rollout. We give you a range on the first call, before any written proposal.