A contract is more than a PDF.
A connected workspace for contracts and obligations, with a draft generator, a knowledge graph and a multi-retrieval system.
Explore the caseThree years building with AI: first inside a large consultancy, now for my own clients. Before that, years on the business side, which is why I start from the problem and not from the model.
I’m the person who finds the right problem,
builds the system and helps your team make it work.
Start with the work that costs your team time and attention. Everything I build falls into one of these three areas, each with real examples.
Information buried in documents.
Connect contracts, files and business knowledge in a workspace built around the way your team thinks.
Every project sits in one of three areas.
Open a case, or jump to the area it belongs to.
A connected workspace for contracts and obligations, with a draft generator, a knowledge graph and a multi-retrieval system.
Explore the caseBringing test design, machine data, charts and written conclusions into a reporting workflow.
Explore the caseAn agentic workflow that ingests and summarises large volumes of clinical files so physicians can prepare a second opinion. Delivered with a team of four over seven months.
Turns bespoke joinery orders into bills of materials, integrated with SAP and the machine production orders.
Next: Parameterising the rest of the product range so more order types are covered.
AI automation for customer service support at a logistics operator, with +40% productivity across the team. I led delivery with full commercial, risk and contractual responsibility.
Assistants, a second brain and connected workflows for notes, tasks, expenses and the calendar. Everything runs on our own hardware and stays private.
Tracks executive appointments and M&A across Europe’s largest companies, and raises an alert where a legal practice may have a commercial opportunity.
Filters information from 15 Latin American authorities and scores its regulatory impact on the mining and energy sector.
Work directly with the person who scopes, builds and takes responsibility for the delivery.
For teams that see the potential but need to decide where and how to begin.
For a concrete workflow that needs a system designed around it.
For teams that want a hands-on partner as their needs evolve.

Engineer by training. Business experience across Europe and Latin America. Previously a project leader at NTT DATA; now an independent AI consultant based in Zürich.
I work with the people who know the business, then build the systems that help them run it. You work directly with me, from the first conversation through delivery.
The full story, in my CVSpeaking at industry events and universities about applying AI to real business processes.
Biggest event: 300+ people







No. Bring the process that feels slow, fragmented or unnecessarily manual. We can start by mapping it and deciding whether an AI system, a simpler automation or a process change makes sense.
Integration is part of the conversation from the start. I look at the tools, data access and constraints you already have before proposing an implementation.
We sign an NDA from the first consultation, or right after the exploratory call, before any business data changes hands. Data access, hosting and tool choices are agreed before implementation. The demos on this site are fictional simulations built on fake data; no client system or document is shown here.
Each project is priced on its scope. As a reference, solutions start from CHF 10,000. We first agree the problem and the deliverables, then the commercial terms, before any work begins.
Yes. I’m based in Zürich and work with international teams in English and Spanish. I take on a limited number of projects at a time, so when we can start depends on my available capacity and what the project needs. We then agree the working rhythm and collaboration format around the project.
We put it into use together, document how it works and agree the support it needs: from occasional adjustments to an ongoing partnership with a shared improvement backlog.
Tell me about the process, the people and where things get stuck. I’ll read it myself.
Discuss your project