I gave this keynote about the different types of AI agents and how to actually wire them up with tools like Make.com, n8n and Zapier. The example I demoed was the one I use myself every morning: an agent that scans my calendar, researches the people I am about to meet, and hands me a short brief before the first call. Nothing exotic, no custom model, just a workflow that removes twenty minutes of prep from every sales day.
How I Choose the Tool, and What Changed Since
The question I get after every keynote is which platform to use. My answer has not changed: match the tool to the token spend, the complexity of the use case, and your compliance obligations. A daily briefing agent does not need an orchestration platform. A multi step workflow touching customer data does, and if you operate in Europe you also need to know where that data goes and how the system is classified under the EU AI Act. Pick the smallest tool that clears both bars.
One update since this talk: for most simple daily workflows you no longer need Make or n8n at all. Claude can handle a large share of the recurring desk work directly, which removes an entire integration layer and its cost. Anthropic documents the agent and tool use patterns in its official documentation, and I would start there before adding any orchestration platform.
If you want help mapping which of your workflows are agent ready and which are not, that is exactly what we do in our workshops.



