Teaching AI Skills: Workshops and Courses That Work

Teaching AI can become a credible service when it solves a defined problem for a defined audience. That means less “master artificial intelligence in an afternoon” and more “use AI to draft safer client updates without leaking confidential data”. Specific beats spectacular.

This guide covers workshops, short courses and team training. For adjacent business models, explore Teaching and consulting or the wider Make Money With AI hub.

Tickd is an independent resource. Nothing here is financial advice or a promise of income. Results vary widely and many people earn nothing. Do your own research.

What's realistic

AI training is not automatically profitable just because the subject is popular. Organisations may have limited budgets, established training suppliers or policies that prevent staff from using public AI tools. Individuals have endless free tutorials competing for their attention.

A realistic starting point is a small, live workshop built around one job. Examples include:

• Using AI to turn rough notes into structured meeting summaries
• Helping recruiters draft clearer job descriptions without automating judgement
• Showing marketing teams how to review AI-generated copy for unsupported claims
• Teaching sole traders to create reusable prompt templates
• Helping managers establish a basic process for checking AI output

The strongest offer is usually practical, narrow and tied to work participants already do. You do not need to know every model or predict where the industry will be in five years. You do need to understand the tools you demonstrate, their limitations and the audience's working environment.

Expect unpaid preparation at first: testing exercises, adapting examples, checking policies and practising delivery. Finding clients can take longer than building the workshop. Repeat bookings are possible when training stays relevant, but they are never assured.

Choose an audience before building the course

“AI for everyone” is not a useful curriculum. A finance administrator, estate agent and charity fundraiser face different tasks, terminology and risks.

Start with an audience you understand through employment, freelancing, volunteering or careful research. The Tickd Industry Playbooks can help you examine how different sectors use AI without pretending every workflow needs automation.

Interview potential learners before recording lessons or making fifty slides. Ask:

• Which repetitive writing, research or administration tasks consume time?
• Which AI tools, if any, are already approved?
• What has the team tried unsuccessfully?
• What information must never be entered into an external tool?
• What would participants need to produce by the end for the session to feel useful?

These conversations often reveal that the real need is not “prompt engineering”. It might be output checking, privacy awareness, workflow design or deciding when not to use AI.

If you want to combine training with deeper operational advice, read How to Become an AI Consultant for One Industry.

Build a workshop around an outcome

A useful beginner workshop should move through explanation, demonstration, guided practice and review. Long lectures about model history rarely change how people work.

For a short live session, a sensible structure is:

• Define one outcome and the boundaries of the session
• Explain what the chosen tool can and cannot reliably do
• Demonstrate a realistic task using non-sensitive sample material
• Let participants complete the task themselves
• Compare outputs and identify errors or weak assumptions
• Provide a checklist or template they can use afterwards
• End with next steps rather than a sales ambush

Create exercises that can survive an awkward demo. AI outputs vary, interfaces change and internet connections misbehave. Keep sample outputs and a text-based backup exercise ready, but be transparent that they are backups rather than live results.

A workshop should also teach verification. Participants need to check facts, sources, tone, permissions and confidential information. If the lesson only shows how quickly a model produces text, it is a product demonstration, not meaningful training.

Workshops or recorded courses?

Live workshops are generally easier to validate because you can observe confusion and answer questions. They also demand facilitation skills, scheduling and repeated delivery.

Recorded courses can serve learners asynchronously, but production creates extra work: scripts, captions, audio, updates and learner support. They are not magically passive. A changed interface can date a screen recording remarkably quickly.

A sensible sequence is:

• Run a pilot live session
• Collect feedback about clarity and relevance
• Revise weak exercises
• Deliver the improved version again
• Record only the material that remains useful across tool updates

Avoid recording a sprawling course before confirming that anyone needs it. The camera may love your slide deck; the market is under no obligation to.

Price the work without making wild claims

Pricing depends on the audience, preparation, customisation, delivery format, group size, usage rights and post-session support. Research comparable professional training in your region, then calculate the time required for discovery, preparation, delivery, administration and updates.

Possible pricing structures include:

• A fixed fee for a defined workshop
• A per-person price for public sessions
• A package covering discovery, delivery and follow-up materials
• A licence for an organisation to reuse course materials internally

Spell out what is included. State whether the client receives a recording, editable slides, templates, follow-up support or permission to reuse your materials. Put cancellation terms and payment dates in writing.

Do not justify a fee with invented productivity gains. If a client wants evidence of impact, agree on modest measures such as attendance, task completion, learner confidence or adoption of an approved checklist. Revenue claims and vague “ten times faster” promises belong in the hype bin. For warning signs, see “Make $10K a Month With AI” Claims: Spot the Hype.

Find clients without pretending to be a guru

Begin with organisations and communities where you understand the work. Professional associations, local business groups, charities, co-working spaces and former colleagues may all reveal genuine training needs. Follow their rules and avoid mass unsolicited messages.

A simple proposal should state:

• Who the session is for
• The problem it addresses
• What participants will practise
• Which tools or accounts are required
• What is outside scope
• How privacy and sensitive information will be handled
• The format, duration, deliverables and fee

Demonstrate competence by publishing a short lesson, checklist or sample exercise. Useful public material builds more trust than calling yourself a visionary. If clients need implementation rather than education, AI Automation for Small Businesses: Where to Start explains the distinction.

Protect learners and your reputation

Before delivery, confirm which tools the organisation permits. Never ask participants to paste customer records, employee information, contracts, health details or unreleased business material into a public model.

Use fictional or properly anonymised exercises. Discuss hallucinations, bias, intellectual property concerns and human review in language suited to the audience. Do not imply that completing your course makes someone legally compliant, professionally certified or qualified to advise others unless a legitimate accreditation explicitly supports that claim.

Keep course materials current and date them. Where features differ by subscription or region, say so. If you use AI to help draft teaching materials, review every claim and example yourself.

Who this is NOT for

This route is a poor fit if you dislike presenting, answering unpredictable questions or revising material frequently. It is also unsuitable if your plan depends on repackaging free tutorials, reading generated slides aloud or claiming expertise after a weekend of experimentation.

Pause before offering training if you cannot explain the limitations of the tools you teach. Knowing where AI fails is part of the job, not an optional disclaimer.

Finally, avoid this model if you expect a recorded course to sell itself. Distribution, support and maintenance remain real work, and many courses earn nothing.

FAQ

Do I need formal teaching qualifications?

Not always, but requirements vary by client and setting. Facilitation experience, subject knowledge and a well-tested lesson matter. Schools, accredited programmes and regulated sectors may require specific qualifications, checks or approvals.

How long should my first AI workshop be?

Keep the first version short enough to test one clear outcome. A focused session with hands-on practice is easier to improve than a full-day survey of every AI topic.

Which AI tool should I teach?

Use a tool your audience can access and is permitted to use. Teach transferable habits—clear instructions, context, verification and privacy—rather than building the entire course around interface buttons.

Should I offer a free pilot?

A limited pilot can generate useful feedback, but define its scope and purpose. Free delivery is not compulsory. You can instead charge a modest test fee, run an internal rehearsal or invite a small group of relevant peers.