You get a team to use AI after training by tying it to real work, not to the course. Name a few champions, give everyone one task they already do every week, share the prompts and Projects that work, have leaders use AI visibly, and measure use in a simple way. Training builds skill; these five things build the habit.
Why usage fades after a good workshop
People leave a workshop motivated, then return to a full inbox. The new skill competes with deadlines, and the old way is faster in the moment because it is familiar. Nobody has decided that AI is now part of how the work gets done, so it stays optional, and optional things lose to urgent ones.
The fix is not more reminders. It is to make the first use easy, relevant and visible. If you are still deciding what training to buy, our guide to AI training in Malaysia covers the options. This article is about what happens in the weeks afterwards.
1. Appoint champions in each team
A champion is a colleague who is a little ahead of the others and is willing to help. They are not a new job title and they do not need to be the most senior person. Their role is simple:
- Answer "how do I do this?" questions in the moment.
- Share one useful example a week in the team chat.
- Report back to whoever owns the rollout on what is blocking people.
Pick people who attended the training and enjoyed it. Give them protected time, even an hour a week, otherwise the role quietly becomes unpaid overtime. For larger organisations, see how this scales in how to roll out AI to a team of 50.
2. Start from real tasks, not features
"Try using AI more" does not work. "Use AI to draft the first version of your weekly report" does. Ask each person to pick one recurring task that is text-heavy, repetitive and low-risk, such as:
- Summarising meeting notes into action items.
- Drafting a first reply to a common customer question.
- Turning a rough outline into a structured document.
- Checking a draft for tone, gaps or unclear sentences.
Keep a human reviewing the result. Once one task works, people find the next on their own. A task that saves even a little time each week is more persuasive than any presentation.
3. Share what works in Projects and a prompt library
The biggest waste after training is everyone rediscovering the same prompts alone. Put the good ones in one place: a shared document, or a shared Project in your AI tool that holds the instructions and reference material for a recurring job. New joiners then start from something that already works rather than a blank page.
Keep the library small and current. Five prompts people use beat fifty nobody opens. Each entry should say what it is for, what input it needs and who reviews the output. Whether shared Projects are available depends on the plan and tool you use, so check what your organisation has before you design around them.
4. Let leaders use AI where people can see it
If managers never mention AI, teams conclude it is not really expected. Leaders do not need to be experts. They need to show ordinary use: "I used Claude to draft this agenda, then edited it", said in a meeting. It also helps when leaders:
- Say plainly which uses are encouraged and which are not.
- Ask "did AI help here?" when reviewing work, without demanding it.
- Accept that early outputs will be imperfect and need editing.
For senior teams who want to set this direction together, the AI Forward CXO Program is built for that conversation.
5. Measure simply, and look at use before results
Do not wait for a perfect return-on-investment model. In the first months, three light questions are enough:
- How many people used AI for work this week? A quick poll will do.
- Which tasks are they using it for?
- What is stopping the others?
The third question matters most. Common blockers are unclear rules on what data may be pasted in, no access to the tool, or no time to practise. Each is fixable, and fixing them is usually cheaper than another training day. Later you can compare time spent on chosen tasks before and after, using your own numbers.
Set data rules before you push adoption
People hesitate when they are unsure what they may paste into an AI tool. Write a one-page guide: what is fine, what needs approval, what is never allowed. Involve your IT and compliance teams, and confirm legal points with them; this article is general information and not legal advice.
When to bring in more structured help
If adoption is patchy after a month or two, a refresher built around your own workflows often works better than a generic course. Agmo Studio, a Select Partner of the Anthropic Claude Partner Network, runs in-house training scoped to your teams and processes. For people who want a recognised credential, the Claude Certified Associate workshop prepares them for the Anthropic exam. Eligible training may be claimable through HRD Corp, subject to eligibility and HRD Corp approval.