Skip to content
Godwin LawsonBUILDING WITH PURPOSE

Training & Workshops

What Should Staff Be Able to Do After a Practical AI Workshop?

What Should Staff Be Able to Do After a Practical AI Workshop?

The practical answer

Staff should be able to choose an appropriate task, provide clear instructions and approved context, inspect the output, verify important facts and explain when human judgement is required. They should also understand the organisation's rules for data, accounts and publishing.

An AI workshop can create excitement quickly. A demonstration produces a draft in seconds, and the audience sees a tool that appears capable of doing almost anything.

The useful question comes afterwards: what can staff now do reliably, within the organisation’s rules, without treating every generated answer as correct?

Practical AI workshop outcomes should connect the tools to defined work and a review method. My corporate training and workshops can be scoped around those tasks rather than a collection of impressive demonstrations.

Choose tasks that are appropriate for practice

Start with work where the team can check the result. Examples include drafting from approved facts, organising non-sensitive notes, proposing an outline or summarising material the reviewer understands.

Avoid beginning with decisions whose consequences participants are not qualified to assess. A short workshop does not make a general AI tool a substitute for professional judgement in every domain.

Ask what part of the task the tool should support and what remains the person’s responsibility. That division should be understandable before the exercise begins.

Give clear instructions and useful context

Staff should be able to state the task, audience, source material, constraints and desired format. A vague instruction often leaves the system to fill important gaps.

Use approved information rather than expecting the tool to know the organisation’s current prices, dates, policies or programme details. Specify which facts must be preserved and which questions remain unresolved.

Participants should also know how to revise the request when the first output is unsuitable. The skill is to diagnose the gap, not simply repeat “make it better” until the wording sounds confident.

Inspect the output before using it

NIST’s generative AI risk guidance identifies the risk of confidently presented false content. Staff therefore need a checking method that goes beyond whether an answer reads smoothly.

Compare the output with the supplied material. Look for missing conditions, invented names, altered numbers, unsupported claims and conclusions that go beyond the evidence.

Ask participants to explain why the output is acceptable or what they changed. A workshop should not reward copying an attractive response without understanding it.

Capability A practical demonstration
Task selection Explain why a chosen task is suitable for AI assistance
Clear instruction Supply the audience, approved facts and required format
Fact checking Identify and correct unsupported or altered details
Judgement Explain what needs human review or another specialist
Responsible handling Use only information and accounts appropriate to the exercise
Repeatability Document a method another staff member can understand

The exact outcomes should fit the team’s role and starting ability.

Apply the organisation’s information rules

Participants should know what information may be entered into a tool and which accounts or services the organisation permits. Do not use real applicant records, private staff information or confidential client documents casually in a demonstration.

Use synthetic or approved material for practice. Explain how to recognise information that needs another route or explicit internal authorisation.

Tool settings and service terms vary. Staff should know where the organisation’s current guidance is kept and who can answer a question, rather than assuming every AI service handles information in the same way.

Preserve the human approval step

Define who approves an output before it is published, sent or used for an important decision. A tool producing a draft does not settle authority to act on it.

For customer-facing work, the reviewer should check the actual offer and commitments. For internal work, they should check whether the output fairly represents the underlying material and its uncertainty.

Avoid teaching a workflow that sends or publishes automatically before participants understand the quality and approval requirements. More automation is not automatically the next learning objective.

Build a reusable task guide

Ask participants to document one approved use case, the information required, an example instruction and a short checking list. Record where the method is useful and where it should stop.

The guide should remain understandable when a tool’s interface changes. Focus on the task and judgement rather than memorising a particular button position.

Keep examples and account guidance current through the organisation’s normal review process. A workshop handout can become outdated if it describes tools or policies as permanent.

Assess a complete attempt

Use a practical task with known source material and an agreed standard. Assess the instruction, output review and explanation, not only the final document’s appearance.

Include a limitation or ambiguity that participants must recognise. They should be able to ask for clarification or preserve uncertainty rather than invent an answer to complete the exercise.

After the workshop, review application to suitable work. Identify where staff need more practice or where the organisation needs clearer rules and templates.

Frequently asked questions

Should a workshop cover every AI tool?

No. Choose tools and tasks relevant to the organisation, with enough time for practice and checking.

Does completing a workshop make staff AI experts?

No. Describe the actual learning and demonstrated tasks accurately. Expertise requires more than attendance at a short session.

What should participants leave with?

An appropriately scoped practical method, evidence of an attempted task, review criteria and a clear understanding of the support or approval needed afterwards.

Measure useful judgement as well as speed

An effective workshop should help staff use a tool deliberately and review its work. Speed matters when the result remains accurate, appropriate and subject to the right human decisions.

Further reading

Reference addresses are provided as plain text for copying.

  • https://www.nist.gov/publications/artificial-intelligence-risk-management-framework-generative-artificial-intelligence

Discuss your next step

Planning an AI workshop for your staff? Share the tasks they handle, the information involved and the outcomes you want. I can help scope practical exercises and review checks around your organisation's needs.

Explore corporate training and workshops

  1. Start on WhatsApp
  2. Send an enquiry through the contact page
  3. Email [email protected]
  4. Call +234 703 799 7791
Lawson Godwin Hezekiah

About Lawson Godwin Hezekiah

Technology entrepreneur, website developer and founder of Mita School and Techmita. I work with businesses, educational institutions, government organisations and NGOs.

Read my story ↗

Keep exploring

Let us work together

Put the ideas into practice.

If your website or digital strategy needs attention, let us discuss a useful next step.