About AI Tools Academy

AI Tools Academy teaches practical AI skills for work, free. It covers ChatGPT, Microsoft Copilot, Claude, Gemini and Power Automate in 100 lessons, assumes no technical background, explains every term the first time it comes up, and is honest about what these tools get wrong. Free role guides and practice exercises help you apply those skills at work.

Why it exists

Using an assistant at work raises practical questions: what information can you share, how should you describe the task, and how do you check the result? The lessons address those questions using fictional workplace material.

Who it's for

Office, admin and professional staff, managers, teachers, apprentices and anyone whose job involves writing, summarising, organising or answering questions. The examples are set in UK workplaces. Specialist courses also cover coding, retrieval and local models, with prerequisites listed on each course.

Who makes it

AI Tools Academy is written and maintained independently in the UK. It is not made by, sponsored by or affiliated with any of the companies whose tools it covers, and nobody pays for a mention in it. It is published without a personal byline: judge it on whether the lessons match what you see on screen, on the dates each lesson shows, and on the public record of what has changed.

How lessons are made

Lessons combine explanations, worked examples and exercises. The main curriculum uses the fictional company Fernway; specialist courses include other invented scenarios. Exercises can be completed without sharing personal or confidential records.

AI tools are used to help draft and check material, the same way the course teaches you to use them. Nothing is published on the strength of an AI answer: claims about how a product behaves are checked against the vendor's own documentation and, where possible, the live product.

How product claims are checked

Product details come from official sources: OpenAI, Microsoft, Anthropic and Google help pages and release notes. Where a detail moves often, like a price, a usage limit or an exact menu label, lessons point to the official page instead of freezing a number that will go stale. Where a feature depends on your plan, your organisation's settings or a gradual rollout, the lesson says so.

How often material is reviewed

Different kinds of material go stale at different speeds, so they are reviewed on different schedules:

• Interface walkthroughs, model pickers, plan and feature availability: about every two to three weeks.
• Product features: about every month.
• Privacy, safety and data guidance: about every two to three months.
• Conceptual, evergreen material: about every six months.
• News and recent changes (event-driven, not scheduled): as things happen.

Every lesson shows the date it was last checked. A date only changes when the lesson has actually been re-checked, so you'll see a mix of dates rather than a uniform stamp.

How errors are corrected

If you spot something wrong or out of date, tell us. Substantive corrections are made in the lesson, its date is updated, and the change is logged on the What's changed page, which also has an RSS feed.

Why the practice data is fictional

Learning to use AI shouldn't mean practising on real customers or colleagues. The Fernway sample files are invented but deliberately messy, like real work, so every exercise can be done without putting anyone's information at risk.

Free learning and independence

The learning resources currently listed on this site are free. We take no advertising and have no affiliate links. No AI company pays for coverage.

What it collects

No accounts, no tracking scripts and no cookies. Your progress and saved prompts stay in your own browser. The full picture is on the privacy page.