After the AI Demo, Can You Do It Yourself?

The important part starts when you take over

I like showing people what is possible. It gives you something to aim at. But watching my screen and using the tool yourself are different experiences. When you take over, you discover the question you did not know to ask: Where did my source go? Why did it add that step? What should I check?

In the workshop, I explain my approach as giving people a direction, then helping them through the obstacles they encounter. Near the end, I ask participants to create a notebook, add some information and use it in a new conversation. The point of that exercise is to find out whether they can repeat the process themselves. The recording does not establish that every participant succeeded.

If you are planning AI training for a school, college or staff team, start there. Decide what someone should be able to do after the demonstration ends.

Try the process yourself, then ask for help

What to notice: Joe asks learners to create a notebook, add something, and use it from a new chat. The check is whether they can perform the process; the recording does not prove every learner succeeded.

1:04. From The One-Person AI Business, 1:01:10.11–1:02:13.96.

Visual context: Joe gives learners an independent practice task while the shared browser remains visible. The instructions are spoken.

Choose an outcome you can actually see

Understanding AI is a large ambition for one session. Turning a short, approved procedure into a checked preparation list is a task you can observe. It gives the learner something useful and gives the instructor something specific to respond to.

For a first workshop, I would use these proposed outcomes:

• Identify which source and version the task uses.
• Produce a short draft for a named reader.
• Find an unsupported addition or confirm that a required detail was preserved.
• Correct the draft and explain the reason for the correction.
• Repeat the process with a different short source, using a reference guide.

This follows a straightforward planning principle: the activity and assessment should match the intended learning. Carnegie Mellon’s Eberly Center explains that relationship. The agenda and exercise below are proposed applications, not measured results from my workshop.

Give everyone a result they can check

Ask learners to create a preparation checklist for a first-time volunteer. They can paste this short source into the approved tool rather than spend the exercise searching for a document.

Using only the practice source above, draft a preparation checklist for a first-time volunteer. Keep required actions separate from details that need confirmation. Do not invent times, contacts or procedures. After the checklist, identify the source sentence supporting each required action.

Suppose a draft tells volunteers to arrive at 9 a.m., put the attendance sheet on the desk and ask visitors to describe their access needs. That is an authored example of a faulty response, not a recorded tool result. It gives the group three concrete things to inspect.

The source specifies door-opening time, not volunteer arrival time. The coordinator keeps the attendance sheet. Questions about access arrangements go to the coordinator; the brief does not instruct volunteers to collect personal details.

Now ask the learner to explain one correction. That explanation tells you more about their review process than a polished checklist alone.

Leave room for practice and a second attempt

Here is a proposed 60-minute agenda for a small introductory session. It is a planning example, not a description of a booked CloudWise service or a promise that every group needs the same time.

1. First 5 minutes: show the finished checklist and define the task.
2. Next 10 minutes: demonstrate one source-to-draft attempt, including a visible source check.
3. Next 15 minutes: let participants make their own checklist and record a question or correction.
4. Next 10 minutes: compare a few decisions and work through access or interpretation problems.
5. Next 15 minutes: repeat with a second short source and less instructor direction.
6. Last 5 minutes: save the work, the reference guide and one next practice task.

Prepare tool access before the session where possible. Have a plain-text source and an example response available if generation fails. Reviewing a supplied response still practices judgment; record that the learner has not yet demonstrated independent tool use.

Check the work and the reasoning separately

Use three simple notes for each outcome: demonstrated independently, demonstrated with support, or not yet demonstrated. A learner may make a correct checklist with help and still need another attempt before they can repeat the task alone.

Keep the evidence modest: the source, the checked draft, one correction with its reason, and the second attempt. A quick explanation such as “I removed 9 a.m. because the source gives only the door-opening time” shows what the learner noticed. If no error appeared, ask them to point to the support for one important detail.

A satisfaction survey answers whether people liked the session. It does not answer whether they can complete this task. If you want to know what carried over, invite a later example using another permitted source. Do not claim that transfer happened before you have evidence.

Plan one useful task before planning the slides

Write one sentence: “By the end, participants will be able to…” Finish it with an action and an observable result. Then choose a short source, a checked example and a second practice task that match.

Your take-home guide can fit on one page: the goal, the practice source, a starter request, the review questions and where to ask for help. For the conversation that starts the exercise, use Start With a Better Conversation. For the wider sequence, follow the complete workflow.

Questions that come up

About this lesson

Adapted in first person from Joe McKenna’s workshop explanations using the automatic timestamped transcript. No direct quotation or audio-verified claim is made. The fictional event source, response, checklist, 60-minute agenda, outcome criteria and evaluation suggestions were authored for this article. Independent practice is requested in the recording; universal participant success or a school implementation is not established.

Source passages: 03:35–04:04; 10:24–10:54; 13:17–13:31; 30:15–30:38; 1:01:02–1:02:16. Source: The One-Person AI Business, September 29, 2026.

Official guidance checked September 30, 2026. Product features can change.

• Carnegie Mellon Eberly Center: Align Assessments, Objectives, Instructional Strategies

AI-assisted organization and supplementary examples. Prepared from Joe McKenna’s workshop.