A hands-on session that gets your team genuinely capable with AI, on your own work. Nothing is theoretical. You try every part of it yourself before lunch, then spend the afternoon on your own work. You leave with things running, and with the time savings measured.

A typical training day

Morning: how it works

1. 9:30 to 9:45 15 minutes Introductions Brief introductions, meet Michael, and see how the day will work.

   Everyone briefly introduces their role, Michael introduces himself and his experience, then takes the group through the shape of the day.

2. 9:45 to 9:55 10 minutes You record the session Start the habit of recording important meetings, so AI has the reasoning behind your decisions, not just the final outcome.

   You start an AI note-taker yourself and use the training session as your first example. The transcript captures the discussion, trade-offs and reasons behind what you decide. You keep it as reusable context for future work and learn how to do the same with your meetings after the training.

3. 9:55 to 10:05 10 minutes What good looks like We’re aiming for something closer to Jarvis from Iron Man: AI that knows you, takes action and works across everything.

   Your ideal AI setup knows you and your business, connects to the tools where your work happens, takes action rather than just giving answers, works proactively and learns from your feedback. That is the target you will work towards throughout the day. If ChatGPT or Claude does not feel like this today, it is because you are still doing much of the work around it: supplying the same context, copying information in and moving the answer back out.

4. 10:05 to 10:15 10 minutes How Claude and ChatGPT do the work See how the Claude or ChatGPT desktop app works with your local folder and connected tools to complete a task.

   A quick tour of the task interface. The model provides the intelligence, your local second-brain folder provides the context, and connectors let AI work with your other tools. You also see how to choose a model, use voice, manage permissions, follow its progress and understand your usage limits. You answer a couple of quick quiz questions to check that you can distinguish the model, its context and the tools it can use.

5. 10:15 to 10:25 10 minutes Review your automation list See what you will build this afternoon and how larger ideas have been broken into small, testable automations.

   We review the automation list prepared from your interviews and show how it was ranked by value and effort. Bigger jobs are broken into their smallest useful components. Once one component works reliably, you can expand its scope or chain it together with others. Keep your list in mind during the morning. As you learn each new building block, think about where it could help. In the afternoon, you will work through the list with Michael there to guide you.

6. 10:25 to 10:35 10 minutes Connect your real tools Connect Claude or ChatGPT to the email, calendar, files and other business tools you already use.

   You connect at least one real tool and test what Claude or ChatGPT can access and do with it. This lets AI find current information and take action where your work already happens, instead of relying on you to copy everything into a chat window. You also see what access you are granting and where your approval is still required.

7. 10:35 to 10:45 10 minutes Break
8. 10:45 to 11:15 30 minutes Set up your second brain See what changes when AI has lasting context about you and your business, then make your prepared second brain your own.

   A second brain is simply a local folder containing the information you would otherwise keep re-explaining: how your business works, your preferences, previous work, real examples and feedback. Claude or ChatGPT uses those files as working context, producing more relevant and consistent work from the start. You compare AI with and without that context, learn how the folder and Obsidian work together, make your own updates, and complete a short quiz to check the idea has landed.

9. 11:15 to 11:25 10 minutes Prompting Learn to brief AI like a highly capable new colleague: naturally, clearly and with the context it cannot guess.

   Tell Claude or ChatGPT what you need, provide the relevant context, show examples and explain what good looks like. Ask for options, break larger jobs into stages, and use voice or screenshots when they communicate something faster. If something fails, explain what was wrong and iterate.

10. 11:25 to 12:00 35 minutes The building blocks, part one Build a skill, schedule it and create a live artifact or Site.

    Skills Save a proven way of completing a task so you can use it again.

    Scheduled tasks Run work automatically at a particular time or on a regular rhythm.

    Live artifacts or Sites Turn information and connected data into an interactive dashboard or tool.

    Everyone completes a small shared exercise with each building block.

11. 12:00 to 12:10 10 minutes Break
12. 12:10 to 12:50 40 minutes The building blocks, part two Try sub-agents, browser and computer use, then your platform’s standout tool.

    Sub-agents Divide a larger job into separate pieces that can run in parallel.

    Browser use Ask AI to open a website, inspect what you are viewing or work through a browser-based task.

    Computer use Let AI operate desktop applications and move work between them.

    Claude Design or ChatGPT Voice Practise the standout capability available on your chosen platform.

    You also learn the fallback order: start with a native connector or plugin, look for an MCP if one is available, then use browser or computer control when there is no direct connection. A short scenario-based quiz checks that everyone can choose the right building block for different kinds of work.

13. 12:50 to 1:35 45 minutes Lunch

Afternoon: your own work

1. 1:35 to 2:50 1 hour 15 minutes Build, test and improve Take one priority from your list through the complete cycle.

   Build it, test it on real work, refine it until it is usable, then measure what changed. Michael will coach you throughout the afternoon, helping you choose the right building blocks and work through any problems. Use your online automation list to work through each priority, record how it is set up, capture the result and check it off before moving to the next one.

2. 2:50 to 3:00 10 minutes Break
3. 3:00 to 4:00 1 hour Continue through your list Complete the same cycle with your next priorities.

   Record the result from each automation, then move to the next item on your prepared list. Most participants will complete two or three usable automations during the day.

4. 4:00 to 4:10 10 minutes Break
5. 4:10 to 4:30 20 minutes Complete what you’re working on Finish, test and measure your current automation.

   Finish what you are working on rather than starting something new. Make sure it is reliable enough to use in real work and that you know how to keep improving it afterwards.

6. 4:30 to 4:45 15 minutes Show and tell Share one automation and what changed.

   Choose one automation to share with the group. Explain what it does, how you built it and the result you measured when you tested it.

7. 4:45 to 5:00 15 minutes Keep it going Maintain what you built and decide what comes next.

   Set up a weekly review that helps you spot what to improve next. Learn how to maintain what you built, add new priorities to your automation list and get help from Michael during the following three weeks.