Up to 5 people
- Each person: the kit, their agent, a certificate
- Online, with me
A hands-on online workshop with me. Each person builds their own AI agent, live, and leaves using it.
What each person builds
It knows what matters to them, so it brings them the right things first and drafts the way they would.
What happens in the workshop
*Connections depend on your Claude plan and your company’s IT policy.
A fair question
Most AI rollouts stop at the license. People need to build the habit hands-on.
The team workshops
How it works after checkout: one email with a seat link per person, to forward to your team, and the link to book your workshop date. Each person adds their name, downloads their own kit, and spends 15 minutes beforehand picking a few pieces of their writing.
Enter the company name and VAT ID at checkout. Secure payment by Stripe.
Launch promotion valid until 31 December 2026.
One real task, two answers
Claude, out of the box
Hi [Client Name], I hope you’re doing well! I’m following up on invoice [number], which is now two weeks overdue. Could you please let me know the status of the payment? I’d really appreciate it. Thanks so much!
Claude with their Gerty agent
[Name], invoice [number] (€6,200) was due on [date]. Pay it by [date, three working days out], or tell me by then when it will be paid. If something is blocking it, tell me what and who owns it. I’ll fix my side the same day.
It names a date, offers a way out, and skips the apology. Because that’s how they write.
Who is behind it
For 15 years I’ve helped thousands of people make technology work for them, at Netflix, Google, Booking.com and Miro, where I led our AI Transformation. Now I do it one person at a time, with AI.
An online session with me and your team.
Everyone builds their own agent, live, and leaves using it.
Privacy
What Claude keeps follows Anthropic’s terms and each person’s privacy settings.
Questions
The kit arrives by email right after checkout. They add it to a new Claude chat, type Go, and Claude interviews them from there. Nothing to install.
Fifteen minutes: pick a few emails, proposals or notes that show how they work. The build happens live, in the workshop.
The 1-hour workshop builds and tests everyone’s agent on one real task. The 2-hour workshop adds a second task, tool connections where your IT allows, and the update routine.
A Claude account (Pro recommended: the build is a long conversation and the result lives in a Claude Project; the kit has a fallback if a plan cannot make a Project).
No. They talk to Claude and answer its questions.
The account your company provides for this, if its policy allows a Claude Project with your own writing. If it does not, a personal account, without client-confidential material. Admins of a company workspace may be able to see what is put in it, so check your policy first.
We never receive anyone’s writing. Each person builds in their own account, from material they choose, and your AI policy still applies.
Each person owns theirs and takes it with them, to other AI platforms too. The kit is for the people in your workshop.
In the 2-hour workshop I show how to connect Claude to tools like Gmail, Calendar and Drive. Whether that works depends on your Claude plan and your company’s IT policy, so we only do it where both allow.
A short routine for keeping an agent current. When someone’s work changes, they tell their agent what changed, it proposes edits to its own instructions and knowledge, and they approve them. Nothing changes without them.
Write to tomas@tha.works.
One workshop, an agent for everyone, and a method they keep.