

Open Lab: Automate your most exhausting task
Your most exhausting task probably has a nickname. The Monday report. The copy-paste marathon from your inbox into a spreadsheet, every week, by hand. A machine could do most of it, and almost nobody gets round to building that machine.
71% of EU companies that considered AI but didn’t use it named missing expertise as a reason (Eurostat, 2025). Open Lab #2 gives you AI expertise in one afternoon, on your own work. You walk in with “this takes me 2 hours” and walk out with a working first version in Claude. No code.
We start with the theory nobody explains properly
Chatbot, workflow or agent? A chatbot answers. A workflow runs the same steps every time. An agent decides the steps itself. Most failed automations picked the wrong one.
Plugging Claude in: connectors and MCP, the open standard Anthropic calls a USB-C port for AI. It’s how Claude reads your email, calendar, Drive and your own tools.
Prompts that run without you: instructions, examples and the token habits that keep an automation sharp and cheap.
Then you build.
Who it’s for
Operations, finance, marketing, HR, sales support, team leads and founders. Anyone with a repetitive task they would love to hand over. You already use Claude or Gemini at work. Lab #1 or similar Claude experience helps.
Agenda
15:45 Doors open, coffee, laptops on
16:00 Chatbot, workflow or agent: which one your task needs, and why most automations end up unused
16:15 Plug Claude in: connectors and MCP, live
16:25 Prompts that run without you: instructions, examples, tokens
16:35 Frame your task and map the workflow on paper
16:55 Break
17:05 Build the first version: a skill, a Cowork task across your files or a connected app. Facilitators at the tables
17:55 Test it on a fresh example and fix what breaks
18:05 Make it stick: handover, human checks, GDPR and the EU AI Act before real data
18:15 Show and tell
18:30 Networking drinks until 19:30
You leave with a working automation, a map of the workflow behind it, and a shortlist of what to automate next.
Your trainers
- Anna Oñate: Founder and Innovation Catalyst at AmPhi Labs, MIT-trained in Applied AI & Data Science and doing an MBA in AI, Data & Analytics. She translates AI into business problems worth solving.
- Kevin Waller: Computer scientist with an MSc in AI and AI/ML engineer at AmPhi Labs. He builds from agents to Machine Learning pipelines and keeps the token bill down.
Bring
A laptop, a paid Claude plan (needed for Cowork and connectors) and one workflow you want off your plate. Use anonymised or dummy examples: no customer lists, personal data or anything confidential. If you don't have such plan, we can facilitate a Build Lab.
After 18:30
Networking drinks: meet like-minded AI enthusiasts, share challenges and learnings, and grow together.
Where and tickets
Capital C, Singel 542, Amsterdam.
€95 early bird for the first 10 seats, then €145 (incl. VAT). Companies get a VAT invoice.
The final agenda may be adjusted slightly to fit the group.
This lab runs with a minimum of 10 participants. If we don’t reach it, we’ll let you know at least 5 working days before and you can move your ticket to the next date or get a full refund.