How do I get my team to use AI?

Give staff a named tool, a specific task, training for their role and permission to use it. Evidence points to training and employer encouragement as the factors tied to use. In the Australian Government's Copilot trial, more training meant more use. Two thirds of participants still used the tool a few times a week or less.

What does the evidence say about adoption inside organisations?

Training and customised training predicted use in the largest Australian trial. iTnews reported that the Digital Transformation Agency's evaluation of the 2024 Microsoft 365 Copilot trial found a correlation between training and usage. Results were better where training was customised to the Australian Public Service context. Focus groups noted participants often forgot Copilot was embedded in the applications.

Two thirds of participants used Copilot a few times a week or less, according to the same report.

A Danish study of 25,000 workers found most employers encouraged AI chatbot use.

Why do staff hold back?

The reasons staff hold back are not measured in the sources checked here. The National AI Centre measured owners. It found around 65% of non-adopting SMEs cite distrust in AI decision-making or a preference for human control, and 54% say AI is not relevant to their business. The Copilot trial found some workers raised concerns about jobs and the environment, according to an Information Age report.

Ask your own team. A short anonymous survey of three questions gives you the answer for your business: What would you use AI for? What worries you? What would you need to feel confident?

What is the rollout sequence?

Seven steps move a team from a first test to a habit. The sequence is Clockwise's own and is not drawn from a study.

1. Choose one task. Use the task scoring method. One task for the whole team avoids confusion.
2. Write the rules first. Approved tools, what data staff may enter, who checks output and who signs off. See the free government AI policy template. The template expects staff to complete required training and report AI incidents.
3. Name a champion. One person who tries the task first and answers questions. In a small business this can be you.
4. Train on the real task. Run a 30 to 60 minute session using your own documents, with a prompt that works. The Copilot trial found customised training gave better results.
5. Set the checking rule. A named person reviews AI output before it reaches a client. The National AI Centre found about half of AI-using SMEs have this safeguard.
6. Measure. Time the task with and without AI for two weeks. See how many hours AI saves.
7. Review monthly. Ask what worked, what failed and what staff want next. Update the prompt and the rules.

What do you say to staff worried about their jobs?

Answer the concern directly and say what you do and do not know. The Danish study found no effects larger than 1% on earnings or hours in 2023 and 2024 data, which the authors state they can rule out. It covers a short period in Denmark. State the rule that fits your business: AI is for drafting and checking, and a person owns every result. Do not promise more than you know.

What mistakes should you avoid?

Avoid four things that stall adoption.

• Buying licences for everyone before you have chosen a task. The Copilot trial found low frequency of use even with licences issued.
• Leaving rules unwritten. Staff will guess, or avoid the tool.
• Skipping the checking rule. Errors reach clients and trust drops.
• Measuring nothing. Without numbers you cannot defend the time staff spend learning.

What to do next

Pick the one task and write the rules this week. The Clockwise Intensive is one focused hour with Megan, one-on-one on Google Meet, on where AI fits in your business and which process to hand over first. It costs $550 AUD.

Book the Clockwise Intensive →

Frequently asked questions

Does training increase AI use among staff?

In the Australian Government's 2024 Copilot trial, the more training participants completed, the more they used the tool, according to iTnews. Results were better where training was customised to the work context. Train on your own tasks and documents.

Should I require staff to use AI?

The evidence checked here does not test requirements. The Danish study found most employers encouraged use. Start with one task, written rules, training and a named reviewer, then ask staff what works before you set expectations.

Editorial standards

Evidence is named and dated. The rollout sequence is Clockwise's own and is labelled as such. This article is general information.

Sources and dates

All checked 29 October 2026.

1. iTnews, Gov Copilot trial unsettled by usage metrics and unmet expectations. October 2024. https://www.itnews.com.au/news/gov-copilot-trial-unsettled-by-usage-and-unmet-expectation-612580
2. Digital Transformation Agency, Evaluation of the whole-of-government trial of Microsoft 365 Copilot. October 2024. https://www.digital.gov.au/initiatives/copilot-trial
3. Information Age (ACS), Govt flags challenges of Microsoft Copilot AI trial. October 2024. https://ia.acs.org.au/article/2024/govt-flags-challenges-of-microsoft-copilot-ai-trial.html
4. National AI Centre, AI adoption insights: December 2025 to February 2026. 7 May 2026. https://www.ai.gov.au/news-and-insights/blog/ai-adoption-insights-december-2025-february-2026
5. Humlum, A. and Vestergaard, E., Large Language Models, Small Labor Market Effects. Becker Friedman Institute Working Paper 2025-56, May 2025. https://bfi.uchicago.edu/wp-content/uploads/2025/04/BFI_WP_2025-56-3.pdf
6. Department of Industry, Science and Resources, AI policy guide and template v1.0. October 2025. https://www.industry.gov.au/sites/default/files/2025-10/ai_policy_guide_and_template.pdf