What a hands-on AI workshop for non-technical staff covers

What a hands-on AI workshop for non-technical staff covers: drafting documents, spreadsheet analysis, and answers from email, using your real work.

My workshops run on your team's own work. Staff open Microsoft Copilot or Claude Cowork, whichever your company uses, and use it on documents, spreadsheets and email from their jobs. By the end of the day, each person has named the tasks they'll hand to AI next week and knows how to check what it produces.

Below is the workshop in the order it runs, from the prep calls to the last hour.

Two weeks before: the prep calls

I talk with the owner or manager and a few staff members. I ask:

• Which tasks take the most hours each week?
• Which ones involve heavy reading, writing, or copying data between systems?
• Which tools do you use: Microsoft 365, Google Workspace, a CRM, an accounting system?
• What information is sensitive and must stay out of AI tools?

The answers become the exercises. An office manager who builds a Monday report from three spreadsheets gets that report as an exercise. A sales team that rewrites the same proposal intro twenty times a month practices on that intro.

First 30 minutes: ground rules

This section is short and has no theory. We cover three rules:

• It can be confidently wrong. You are still responsible for anything you send.
• It only knows what you give it. Leave out the right document and it will fill gaps with guesses.
• Some information stays out. We go through your company's rules for customer data, financials and confidential material.

Drafting documents

Exercises:

• Outline to draft. Paste in bullet points, ask for a first draft, edit it.
• New audience. Turn a technical explanation into a customer email, or a long memo into a five-line update.
• From source material. Point the tool at an old proposal, a policy or meeting notes and ask for a new document based on it.
• Specific requests. We compare "write a follow-up email" with "write a short follow-up to a customer who asked about lead times, friendly tone, under 100 words, say we'll confirm by Friday," and look at both results side by side.

A typical exercise: an account manager at a manufacturer has to tell a customer an order is late. She pastes in the customer's email and the production team's update, then asks for a reply that apologizes, gives a new date, and doesn't promise more than production can deliver. She edits a couple of sentences and sends it. She can reuse the same steps for the next delay notice.

Spreadsheet analysis

Many non-technical staff use Excel less confidently than they write, so this section moves slowly. Exercises:

• Plain-English questions. "Which customers ordered less this quarter than last?" "What's the average days-to-pay by customer?"
• Formulas. Describe the result you want and have the tool write or explain the formula.
• Report summaries. Take an accounting or CRM export and ask for the three points a manager should see.
• Checking. Recalculate a few numbers by hand every time until you've seen the tool get that type of task right repeatedly.

Copilot does this inside Excel. Claude Cowork reads a file or folder and builds a summary or a new spreadsheet, which helps when the data is spread across several files.

Email and other apps

What staff practice here depends on the tool and which apps are connected:

• Thread summaries. "What did we agree to with this vendor, and what's still open?"
• Missed requests. "List this week's customer requests that haven't had a reply."
• Meeting prep. Gather recent emails, notes and files about a client into a one-page brief.

Copilot works across Outlook, Teams and your Microsoft files. Claude Cowork connects to Gmail, Google Drive, Slack and your calendar. We connect only the apps the team needs.

Last hour: each person's list

Each person writes down two or three tasks from their job that they will hand to AI in the next week, such as "Monday sales recap" or "first drafts of quote follow-ups." We run at least one of them together before the session ends, so nobody leaves without a working example on their own material.

Outside the workshop's scope

A workshop teaches people to use an AI tool in their daily work. Automated workflows that run without a person, connections between your systems, and process redesign are a separate project. If they come up during the day, we discuss them as AI implementation.