AI Enablement Program: A 90-Day Plan for Policy, Tools and Training

Most companies with 50 to 500 staff already use AI. The question is whether they use it on purpose. Some people paste client text into a personal chatbot account. Others wait for permission that never comes. An AI enablement program closes that gap. This guide gives you the parts of a program, the order to build them in, and a 90-day plan with owners and outputs.

What is AI enablement, and how is it different from AI training?

AI enablement is the work that makes AI use safe, useful and normal across a company. Training is one part of it. A program also sets the rules, chooses the tools, names the people who help others, and measures what changes.

Training alone often fails. In the BCG AI at Work 2025 survey, regular use of generative AI among frontline employees “has stalled at 51%”, and only one-third say they were properly trained. People need clear rules and good tools before a course makes sense.

A full program has five parts:

• Policy. What staff may do, with which data, and who checks the output.
• Tools. Approved accounts with the right data settings.
• Training. Short, task-based sessions for each team.
• Support. A few named people who answer questions in each team.
• Measurement. Use, quality and time saved, tracked over time.

Does the EU AI Act require an AI enablement program?

The AI Act does not ask for a “program” by name. It does ask providers and deployers of AI systems to support AI literacy. Since the Digital Omnibus on AI changed the text, Article 4 says they “shall take measures to support the development of AI literacy” of their staff. It also says this “does not require providers or deployers to guarantee any specific level of AI literacy of any individual.”

Three points from the Commission’s AI literacy Q&A matter for planning:

• Article 4 applies since 2 February 2025. Supervision and enforcement rules apply from 2 August 2026.
• There is no duty to test staff, and “there is no need for a certificate”. Internal training records are enough.
• Staff who use ChatGPT for tasks such as copy or translation are in scope. They should know the specific risks, for example hallucination.

A company outside the EU can be covered too. Article 2(1)(c) includes providers and deployers in a third country where the AI output is used in the Union. If you sell into the EU, read Article 2 with your lawyer to check your position. This article is not legal advice.

What should the AI usage policy say?

The policy says which tools staff may use, which data they may put in, and who reviews the output. Keep it to two pages so people read it.

Cover these points:

1. Approved tools. List them by name. Everything else needs approval.
2. Data rules. No personal data, client secrets or contracts in tools without a company agreement.
3. Human review. A named person checks any output that goes to a client, a contract or a public page.
4. Disclosure. When staff must say that AI helped.
5. Questions. Who to ask, and how to report a mistake.

Write it in the language your staff use every day. If your company works in both Latin and Cyrillic script, publish both.

Which tools should staff get, and how do you stop shadow AI?

Give staff approved tools that are good enough for their real tasks. A ban without an alternative does not work. BCG found that more than half of employees will find and use other tools when they lack approved ones. The Microsoft Work Trend Index 2024 found that 78% of AI users bring their own AI tools to work.

Start with one general assistant on a business plan, plus the AI features already inside the software you pay for. Then add tools for specific tasks when a team shows a need. When a task needs AI inside your own systems, our AI automation service builds that workflow with a review step where a wrong answer costs money.

How much training is enough?

Plan for at least five hours per person, with in-person coaching. BCG found that regular use is much higher when staff get at least five hours of training plus coaching. Short videos alone do not reach that level.

Make the sessions about real tasks. A sales team practises a proposal draft. A finance team practises checking a summary against the source. Every session covers the policy and the risk of confident wrong answers. Keep a simple record of who attended and what the session covered. That record is your Article 4 evidence.

What do AI champions do?

AI champions are staff who use AI well and help their colleagues. They answer questions, share prompts that work and report problems to the program owner. Pick one per team to start. Choose people others already ask for help, not only the most technical.

Managers matter as much as champions. BCG found that strong leadership support raises the share of employees who feel positive about generative AI from 15% to 55%. Only about a quarter of frontline staff get that support.

How do you measure adoption beyond licence counts?

Measure what people do with the tools, not how many accounts you bought. Licence counts hide the gap between leaders and staff. In the Microsoft Work Trend Index 2025, 69% of leaders use generative AI at least several times a week, against 45% of employees.

Track a small set each month:

• Active use. Share of staff who used an approved tool each week.
• Tasks. Which tasks each team uses AI for.
• Quality. Errors found in review.
• Time. Minutes saved on two or three tasks you timed before the start.
• Clarity. A short survey question: “Do you know what you may use AI for?”

The last one matters. Gallup found that 47% of US employees say their organisation has integrated AI tools, but only 25% say it has shared a clear AI plan.

What does a 90-day plan look like?

A 90-day plan builds the parts in order: policy first, then tools, then training and measurement. Use this table as a starting point.

If a pilot from week 13 needs to become a daily process, read our guide on moving an AI pilot to production. If you run a smaller team, the small business version is shorter.

Where do you start?

Start with the survey in weeks 1 and 2. It shows you which tools people already use and which tasks matter. If you want help to plan the program or connect it to a wider digital strategy, book a call with us. We are a team of 5 in Sarajevo, with clients in Bosnia, the EU and the US.

Sources

• EU AI Act Service Desk, Article 4: AI literacy — consolidated text as at 27 July 2026
• EU AI Act Service Desk, Article 2: Scope — consolidated text as at 27 July 2026
• European Commission, AI Literacy questions and answers — last updated 20 October 2026
• BCG, AI at Work 2025: Momentum builds, but gaps remain — 26 June 2025
• Microsoft, Work Trend Index 2025: The year the Frontier Firm is born — 23 April 2025
• Microsoft, Work Trend Index 2024: AI at work is here, now comes the hard part — 8 May 2024
• Gallup, Artificial intelligence indicator — figures as of May 2026