Present AI. Sell AI. Build AI.

Presenting Microsoft AI

Executives don’t buy models. They buy a story they can repeat to their board. I turn AI into that story: one idea, one number, one slide.

Here is one I present. Most large companies pay consultants millions to explain their own business back to them. An ontology in Microsoft Fabric maps that business once (customers, products, contracts, costs, and how they connect), and AI agents answer the questions on top of it.

Selling Microsoft AI

A good story opens the door. A number closes the deal. I sell AI on outcomes, using Return on AI (ROAI): the annual value an AI solution creates, divided by what it costs to run.

For the ontology case, the hypothesis is simple. A company spends CHF 10m a year on management consultants. If 80% of that analysis is addressable by ontology AI, that is CHF 8m in value. At CHF 1.6m in annual cost, the ROAI is 5x. I test that hypothesis with the CFO in a discovery workshop, then put it in the proposal.

Building Microsoft AI

Then I deliver what I sold. The ontology is built in Microsoft Fabric, on governed enterprise data in the lakehouse. In Microsoft Foundry, multiple AI agents work on top of it, each with its own domain, such as finance, operations and sales, with an orchestrator that routes every question to the right one.

It starts as a working prototype on one business question and goes to production with security, monitoring, evaluation and cost control built in from day one.

Swiss native, based in Zurich

Languages: Swiss German (native) · German (native) · English (fluent)