AI Automation
AI Training for Employees: What to Cover and What It Costs
Train for the real work. AI training for employees works best as one shared half-day on the team’s own tasks, followed by role-specific depth for managers, people who handle personal data and the few who will build workflows. Our half-day workshops start from £1,500 and company-wide programmes from £3,000, both on our AI literacy workshops page. In the EU there is a legal duty to support staff AI literacy, but it does not require a certificate or a fixed number of hours.
Most courses you will find are generic: a tour of ChatGPT, some slides and a quiz. Training works better when it uses the tasks your people already do and the tools your company already approves. This article covers what staff should learn, what each role needs, what a good workshop looks like (with one real example), what it costs, what the law says, and how to tell whether it worked.
Key takeaways
- Everyone needs the same five basics: briefing an AI tool, checking its output, knowing what data never goes in, knowing when not to use AI, and knowing which tools are approved.
- Add depth by role. Managers, people who handle customer or personal data, and workflow builders each need something the rest of the team does not.
- A half-day on your own tasks beats a generic course. Follow it with a short refresher, because habits drift once the session ends.
- Our team training starts from £1,500 (up to 15 people) and company-wide programmes from £3,000, fixed and agreed before work starts.
- The EU AI Act asks providers and deployers to support staff AI literacy. It imposes no mandatory course or certificate, and it can reach UK and US firms. This is not legal advice.
What should AI training for staff include?
Good AI training for staff teaches five skills, and none of them is a clever prompt trick. They are a way of working that people can repeat on Monday morning.
Writing a brief. An AI tool gives you a first draft of whatever you describe, so the quality of the draft follows the quality of the description. People learn to give the audience, the purpose, the constraints, an example of the tone they want and the thing to avoid. A reusable structure beats a collection of magic phrases.
Checking the output. AI tools state wrong things in the same confident voice as right things. The European Commission’s AI literacy Q&A makes this point with the example of staff using ChatGPT to write advertising text or translate. It says they “should be informed about the specific risks, for example hallucination” (European Commission). In practice, people need a habit: check names, numbers, quotations and claims against the source before anything leaves the building.
Knowing what data goes where. This is the rule with the most downside. A public chatbot account and a business account with a contract behind it are different places, and staff need a plain list of what can go into each. Client data, contracts, HR records and anything you would not email to a stranger stay out of public tools. Our ChatGPT vs Claude comparison covers where the data goes in the two assistants most teams start with.
Knowing when not to use AI. Some tasks are poor fits: decisions that affect a person’s job or money, anything that needs a source you can cite, and work where one wrong detail is expensive. Training should name those tasks for your business, with examples, so people do not have to guess.
Knowing the approved tools. Staff will use something, with or without a policy. Training goes better when it points at the tools the company has chosen, shows how to log in, and says who to ask when a task does not fit.
What each role needs to learn
Everyone gets the five basics. The table shows what to add on top for each group.
| Group | Adds | Format |
|---|---|---|
| Everyone | The five basics | Half-day, hands-on |
| Managers | Setting rules, reviewing AI work | Short add-on |
| Data handlers | Privacy rules, redaction | Focused session |
| Workflow builders | Tools, testing, handover | Multi-session |
Managers decide what good looks like. They need to know how to review work that AI helped produce, how to ask “what did you check?”, and how to set expectations without banning the tools or waving everything through.
People who handle customer or personal data need the most careful session. In the UK, the Information Commissioner’s Office (ICO) puts the principle plainly in its AI guidance: “The key is that you only process the personal data you need for your purpose” (ICO). That page carries a notice that it is under review because of the Data (Use and Access) Act, so check it before you rely on it. Training for this group covers what counts as personal data, how to strip it from a prompt, and when to stop and ask. The ICO also says AI tools that process or are trained on personal data “require particular attention” (ICO).
The few who will build workflows will connect AI to your spreadsheets, inbox or CRM. They need more time: how to test a workflow on real examples, where a person should approve the result, and how to hand it over. If the workflow becomes business-critical, that is the point to talk about AI automation consulting rather than training.
Should you train everyone or only some staff?
Train everyone on the basics. The risks come from the least-trained person with a browser tab open, and the benefit comes from the whole team working the same way. The Commission says “having different levels of training or learning approaches could be appropriate” for staff with different backgrounds (European Commission), which supports a shared base plus depth for those who need it. It also notes that “other persons” under the duty can include a contractor or service provider, so include them if they use your tools.
What does a good AI workshop for employees look like?
A good AI workshop for employees is hands-on, built on the team’s own work, and ends with something people can reuse. Here is a half-day agenda we would recommend as a starting point, for one team of up to about 15. Times are indicative.
- Where AI helps and where it does not (30 minutes). Start from the team’s real week, not from the technology.
- Briefing an AI tool (60 minutes). Each person writes a brief for one of their own tasks, runs it, and compares results with a neighbour.
- Checking the output (45 minutes). Take AI drafts that contain planted errors and find them. People remember the error they caught themselves.
- Data rules (30 minutes). Walk through a plain list of what goes into which tool, using examples from the team’s own files.
- Use cases for the next month (45 minutes). Each person picks two tasks to try and names one thing they will not hand to AI.
- Wrap-up and the first draft of a policy (30 minutes). Capture the rules the room just agreed while they are fresh.
Our service page lists what a workshop includes. It covers exercises built around your tools and workflows, hands-on work rather than just slides, and what never goes into a public chatbot. You also get a starting point for your AI policy and materials you can reuse afterwards.
A multi-session programme spreads the same material over several weeks. A second session a few weeks later reviews what people actually tried, fixes habits that slipped and adds the role-specific depth. For larger organisations, our company-wide format runs team by team with follow-up clinics.
A worked example: the Dellar Davies workshop
Dellar Davies is an events agency in the travel sector, and a few people there had started using AI tools on their own while the rest of the team had not. The agency asked for a practical introduction to tools it could use straight away, rather than a bespoke build.
We began with a discovery call on how the team spends its week: promotion for events, e-flyers to guest lists, newsletters, LinkedIn posts, pitches to new clients and write-ups of post-event feedback. We then ran an in-person workshop at the agency’s team day, built around that content rather than generic examples. The team saw off-the-shelf writing tools such as Copy.AI and Hoppy Copy, and image generators, used live on examples from their own sector.
Every example started from a structured prompt covering the brand, the audience, what to cover, the offer and the call to action. The team could then reuse it for their next event and get a first draft instead of a blank page. Nothing was built. The case study has the detail, including the examples we used, on the Dellar Davies case study page.
The order of work is worth copying: find out what people do all week, pick tools that fit it, and give them a reusable structure.
How much does AI training for a team cost?
Prices for AI training vary by what you buy, and the useful comparison is what the price includes. Ours are published and fixed before work starts.
| Format | Price | Fits |
|---|---|---|
| Team training | From £1,500 | Half-day or multi-session, up to 15 |
| Company-wide programme | From £3,000 | Every team, with follow-up |
Both come from our AI literacy workshops page. The company-wide programme runs team by team with follow-up clinics, drafts an AI use policy with you and trains champions to keep it going. Anything outside that is quoted as a custom programme.
A few things move the price with any provider, and they are worth asking about before you compare quotes:
- Group size. Hands-on work needs a room, not a lecture hall, which is why larger organisations run team by team.
- In person or remote. In person suits a team that works in one place and wants the room’s energy. Remote suits spread-out teams, but hands-on exercises need more structure on a screen. Ask whether the price changes by format.
- Customisation. A course built on your tasks takes preparation. A generic course does not.
- Follow-up. One session gives you a spike of interest. A follow-up a few weeks later gives you a habit.
- Ownership of materials. Check you can reuse the prompts, examples and handouts afterwards.
If you are unsure where AI would help your business most, a different first step may suit you better. Our AI readiness assessment starts from £1,500 and looks at your data, workflows and team skills over one week for one department, and training can follow from what it finds. Our AI readiness checklist is a free self-assessment, and AI for small business lists ten jobs to hand to AI first.
Is AI training a legal requirement?
In the EU, yes in a limited sense. Article 4 of the AI Act was replaced by Regulation (EU) 2026/1744, the Digital Omnibus on AI, published in the Official Journal on 24 July 2026. The new text says providers and deployers of AI systems “shall take measures to support the development of AI literacy of their staff” (EUR-Lex). The duty also covers other people dealing with the operation and use of AI systems on their behalf. The same paragraph adds that the obligation “does not require providers or deployers to guarantee any specific level of AI literacy of any individual.”
The Commission says Article 4 entered into application on 2 February 2025, and that the supervision and enforcement rules apply from 2 August 2026 onwards (European Commission). It also says “no strict requirements or mandatory trainings are imposed” and “There is no need for a certificate.” National market surveillance authorities could impose penalties, and any sanction must be proportionate to the case.
Who it reaches. The AI Act covers providers and deployers established in a third country where the output of the AI system is used in the Union (EUR-Lex, Article 2(1)(c)). The Commission adds that the duty applies to actors inside and outside the EU where the use has an impact on people in the EU. A UK or US company that uses AI tools in a way that affects people in the EU can therefore be in scope. Whether your use meets that test is a question for a lawyer, and this article is not legal advice.
What about the UK and US? We did not find a general legal duty in either place that mirrors Article 4, and we have not claimed one. Other rules still apply to what staff do with the tools. In the UK, data protection law still applies to personal data that goes into AI tools. The ICO says organisations using AI on high-risk personal data should have “a data protection impact assessment (DPIA) and appropriate safeguards in place” (ICO). If you work in a regulated sector in either country, check its own rules.
Whatever the rules, written evidence of what you did is useful. A short record of who was trained, on what and when is cheap to keep.
How do you know AI training worked?
Measure behaviour, not attendance. You do not need a dashboard. You need three or four signals that a manager can read in a team meeting.
- Approved tool usage. Are people using the tools you provided, or has usage stalled after the first fortnight?
- Tasks moved. Ask each person which two tasks they now do with AI, and whether the result was good enough. A list of tasks and owners tells you more than a satisfaction survey.
- Fewer policy slips. Track the near misses people report, such as personal data in a prompt or an unchecked claim in a draft. A rising count early on is a good sign, because it means people are speaking up.
- Review quality. Do managers see fewer AI errors reaching customers?
Ask the same questions before the training and about six weeks after it, and write the answers down. The right target depends on your team and your starting point.
Pair the training with a written AI use policy. A policy should exist before staff are expected to follow it. It does not need to be long. Name the approved tools, list what data stays out of each, say which tasks need human review before anything goes out, and name a person to ask. The workshop is a good moment to draft it, because the room has just argued through the edge cases. Our company-wide programme drafts the policy with you for that reason.
A next step for your team
If you want to move from reading to booking, start with the smallest useful version: one team, one half-day, built on their real tasks. Our AI literacy workshops start from £1,500 for up to 15 people, with the price fixed and agreed before work starts. Send us the five tasks your team repeats most often and the tools you already pay for, and we will shape the session around them.
If you would rather pin down where AI pays off before training anyone, the AI readiness assessment is the better first step. If you are unsure which of the two fits, the free 30-minute consultation below is the place to ask.
Frequently asked questions
How long should AI training for employees take?
For most teams, one half-day session on their own tasks gets people using AI tools on real work, and a short follow-up a few weeks later fixes the habits that slipped. Larger organisations usually run sessions team by team. Our half-day workshops start from £1,500 for up to 15 people, and company-wide programmes with follow-up start from £3,000. Length matters less than whether the exercises use your own work.
What should AI training for staff include?
Five things. Staff learn to write a clear brief for an AI tool, check what comes back and know which data never goes into which tool. They also learn when not to use AI, and which approved tools your company actually provides. Managers, people handling customer or personal data, and the few who will build workflows each need extra on top of that shared base.
Is AI training a legal requirement?
In the EU, Article 4 of the AI Act requires providers and deployers of AI systems to take measures to support the AI literacy of their staff. The Commission says no mandatory training or certificate is imposed. It can reach UK and US firms whose AI use affects people in the EU. We found no equivalent general duty in UK or US law. This is not legal advice.
How much does AI training for a team cost?
Our team training starts from £1,500 for a half-day or multi-session format for up to 15 people, and company-wide programmes run team by team with follow-up start from £3,000. Prices are fixed and agreed before work starts. Compare any provider on what the price includes: whether the exercises use your own tools and tasks, whether there is follow-up, and whether you keep the materials.
Should we train everyone or only some staff?
Train everyone on the shared basics, because one person pasting a client contract into a public chatbot is enough to cause harm. Then add depth by role: managers who set expectations, people who handle customer or personal data, and the few who will build workflows. The European Commission says different levels of training can be appropriate for staff with different backgrounds.
Planning AI training for your team?
Book a free 30-minute consultation and we'll shape a workshop around your team's own tools and tasks.