Training your teams on AI: where do you start?

Ask your team which AI tools they use and what they use them for. Then choose a few work documents approved for training. Use them to practise writing a request, checking the result and recognising information that should stay out of the tool.

Why training needs to continue over time

In the BCG AI at Work 2025 survey, only 36% of employees surveyed consider their AI training sufficient. That figure points to the need for better suited training, but it does not prescribe a universal duration.

After the introduction, let the team practise on its own tasks. Review the results together: what needed correcting, what was unclear and where did people need help? Use those answers to plan the next session.

In two US surveys reported by McKinsey in 2025, executives estimated that 4% of employees used generative AI for at least 30% of their daily work. In a separate employee survey, 13% reported that level of use themselves. These are an estimate and a self-report, not a productivity measurement or construction-sector result. The difference suggests checking actual use within a company before designing support.

The two essential dimensions

Set aside time for checking results and handling data.

Start with judgment: recognising an answer that needs checking, finding its source and knowing when to ask for specialist advice. The company should define who approves the result.

Next, look at data: which documents can go into which tool? A personal plan can be free or paid. The service, its terms and its settings determine the relevant conditions. Practise making this distinction with approved examples before copying documents into a tool.

These skills help the team use AI within the company's rules. Make sure people also know whom to ask when a new or unclear case comes up.

3 signs that training holds up in daily work

Use these points to assess a course you have taken or are considering. The precise content depends on the actual tasks of the office and the tools already in place.

• Staff know how to phrase a request and, on their own office tasks, which outputs need careful review.
• The question of what can or cannot be entered into a public tool was raised explicitly, with examples drawn from the company's usual documents.
• There is a clear answer to this question: what happens when the employee who was most skilled with the AI tool leaves the company?

Share knowledge across the team. Keep useful examples and plan how to help new colleagues learn the same practices.

A shared page can record approved tasks, instructions and checks. Add examples and keep them up to date. This makes handovers easier, though a demonstration and supervised practice may still be needed.

The construction site question

Do teams need to leave the construction site to be trained? It depends on actual use. AI literacy concerns people who use these systems for the company, in the office and on site.

In the office, this may include people who handle quotes, site meeting reports, technical specifications and supplier correspondence. On site, others may use AI tools for different tasks. Identifying these uses lets the company match training to the tasks of each group.

What the EU AI Act requires, without interpreting it for your situation

Since 27 July 2026, the amended Article 4 on EUR-Lex has required providers and deployers of AI systems to take measures that support the development of AI literacy among people who operate or use those systems on their behalf. They must consider knowledge, experience, education and training, the context of use, and the people affected. They do not have to guarantee a particular level for each individual. The article sets neither a universal number of hours nor a single training format.

The change was made by Regulation (EU) 2026/1744. It replaced the earlier wording about ensuring a sufficient level of AI literacy to the best of a provider's or deployer's ability.

In practice, recording the content and intended audience of the measures taken makes the approach easier to review. A record alone does not prove any individual's level of AI literacy.

The precise interpretation of Article 4 in your situation belongs in legal advice.

Frequently asked questions

Is a single training session enough?

It depends on the team's experience and intended tasks. An initial session can establish the basics. Use exercises and follow-up in daily work to check whether people can apply the rules and review results.

Do teams need to leave the construction site to be trained?

Not automatically. First identify who actually uses AI systems, in the office and on site. Train those people on their own tasks. Staff who do not use those systems do not all need to leave the site for the same session.

Who does the AI Act obligation apply to?

Article 4, as amended from 27 July 2026, addresses providers and deployers of AI systems. A construction SME that uses such a system in its operations may be a deployer. The obligation concerns measures supporting AI literacy, without a guarantee of a particular level for every individual. Consult a specialist lawyer about your particular situation.

How long does it take to train a team?

There is no universal duration. A few hours on real tasks can establish the basics. Only later use and review can show whether the team applies what it learned over time.

---

To prepare, choose a few recurring tasks, such as a quote or a site meeting report. Identify who handles them and gather approved examples for the exercises.

The tools in use, participants' experience and the problems they encounter will shape the course. Our AI training starts with those uses.