AI team training helps people use approved AI tools on real work, judge what comes back, and know when not to use it. We train on your own tasks and your own stack, by role, with the data and review rules your business needs.
When the licences aren't paying off
If your team has AI tools but no training or shared way of working, a few people benefit and everyone else waits.
People are using AI on personal accounts, and nobody knows what's going into it.
You bought Copilot, ChatGPT or Gemini licences, and usage has flatlined.
Two people are saving hours a week. Everyone else is waiting to be shown how.
Output quality and tone vary depending on who used which tool.
You want automations and agents, but nobody would own them once they're built.
Most teams were never shown how
Training hasn't kept up with the tools, so people teach themselves and reach for apps nobody has approved.
60%
of UK businesses with 500 or more staff say employees haven't completed comprehensive AI training.
68%
of UK businesses with 500 or more staff say employees use unapproved AI tools at least occasionally.
36%
of employees surveyed worldwide feel adequately trained to use AI.
Source: BCG, AI at Work 2025, June 2025, opens external site
The tools and licences you have, your AI policy if there is one, and who uses what today. A short survey and a few interviews.
Hands-on sessions in your tools, on tasks your team actually does, in small groups. On-site or remote.
Prompts, skills and one-page guidelines, organised by task, so good habits don't depend on two enthusiasts.
Built with the people who'll run them, each with a named owner, limits and a log.
We time the tasks we trained on, before and after, and report what changed.
We pick the modules that fit each role. Every session uses your tools, your brand and your real tasks, not generic demos.
What AI is good at, where it gets things wrong, and what should never go into it.
Turn hours of reading into a briefing you can trust.
How AI search tools find and cite brands, and how to use them for research without being misled.
Copy, images and video that sound and look like you.
Take the copy-and-paste out of the week.
AI for the work around the work.
Agents that do real tasks, with an owner, limits and a log.
Tools we teach on: whatever you already pay for. Most teams use some of Claude, ChatGPT, Microsoft Copilot, Gemini, Canva, Adobe Firefly, HubSpot and automation tools like Zapier or Make.
What good training looks like
Three things that make it stick.
The useful skills are choosing the right task, giving AI the right material, checking what comes back, and knowing when a person must decide.
Habits change when people have approved tools, shared examples, someone to ask, and a reason to use them the following week.
Leaders, marketers, operations and risk teams use AI differently. Training by role is quicker and stays relevant.
Where's your team now?
Five quick questions to see where training would help most.
How we teach it
Our AI management system is independently audited. Everything we teach about safe use, data and oversight comes from running it.
We build with Claude in client products and our own tools.
We've built AI tools for our own new business, scoping and delivery work, and trained our team on them before training anyone else's.
The things people usually ask
Plain answers about AI team training, what's involved and what we need from you. Anything else, just ask.
Relay is the fixed-price starting point for one team, in three weeks. This is for larger programmes across several teams or roles, built from the same modules.
It helps, but it isn't required. If you don't have one, we train to sensible defaults and recommend Guard alongside. If you do, every session is built around it.
The ones you already pay for. We don't sell licences, and we'll tell you if one isn't earning its place.
We train in your accounts, under your settings, and check data handling with you before the first session. Nothing goes into a tool your policy doesn't allow.
We time the tasks we trained on before and after, and track how many people are using the tools a month later.
