Let’s leave the likely to the machines. Let’s fall in love with the possible.
I work with companies to find out what their customers need, and develop ideas, brands and products from it. And I work out what machines may do and what people must make.
Book a conversationFor owners and managing directors who have to decide what to develop next, for whom, and what role “AI” should play in it.
What should be made, and why.
It can seem as if AI agents – and soon robots – could capture, plan and build anything. Yet some tasks stay with us. Above all: describing what should be made, for whom and to what end. That takes someone who understands the people it is made for: their situation and goals, their wishes and fears, their problems and workarounds. This has to be turned into requirements against which every result can be measured. Relationships and experiences cannot be cast in code. They have to be intended, thought through, pursued, lived and judged. The work is not running out; it is only shifting.
What must people make?
- Ideas and innovation
- Finding new products, services and lines of business – from observation, conversation and what technology now makes possible. Making ideas tangible as prototypes before serious money is spent. Working out whether customers want them and whether they can be built. And deciding which are worth it.
- Brand and communication
- Name, brand and voice that people recognise. That includes positioning and visual identity: what a company stands for, what sets it apart and how it looks. The result: a design system for imagery, type and copy that applies everywhere. Even when machines do some of the writing.
- Product and design
- Designing what people use: on screen, in physical spaces and in services. Information architecture and interaction, form and material – tested with users early and often. Worked out precisely enough for teams and AI agents to build from. And staying with it until the product is on the market.
- Leadership and teams
- Bringing people along as their work shifts. Reshaping roles and developing the skills needed once AI agents work alongside them. Finding the right people, building and leading teams, on an interim basis too. So that it is clear who is responsible for what, and everyone can work with confidence.
What may machines do?
- Assessment
- Where AI brings real value and where only effort. Looking at processes and data, cost and impact, law and security, the team’s know-how and what to buy versus build. The result: priorities, a roadmap and named risks.
- Adoption
- How AI agents get into products and processes, and who steers them. Consolidating data and making it accessible to machines. Encapsulating the company’s unique capabilities. And keeping the solution open to other models and providers without giving up sovereignty over the company’s own data.
- Governance
- Which rules AI works by, and which person signs off which level of quality. Plus machine controls that also have to be developed first: checks, tests and limits that take effect automatically. Every sign-off checks whether what was meant to be made has been made.
- Operations
- Whether the value materialises and lasts – and what difference customers notice. Measuring time, cost and rework, before and after. New models first have to prove themselves on the same tasks, and when a service goes down, there is a way to work without AI. And regularly deciding what machines may do.
What I’m working on right now
What stays, what goes, what comes first.
For a mid-sized company, I’m guiding the rebuild of an IT landscape that has grown over decades. It is about people and processes as much as architecture and priorities. A small team is building the new system on a shared data foundation while the business keeps running. The team achieves today what a year ago was neither conceivable nor affordable. Part of the work is developing the governance for it.
My own projects and products: I decide what gets made, and why.
Alongside this, I develop my own projects and products. Idea, name, brand, design and requirements are my work. With me, all of this sits in one pair of hands; in teams and organisations, it has to be explicitly governed.
The crafts remain: user research, design, copy, software engineering and project management ensure the quality of their part. Their efficiency can rise dramatically – if review and sign-off grow with it.
- What has to be settled before AI agents build?
- Who it is for, what it should achieve and how you can tell it has succeeded.
- What changes when a team works on it?
- One person leads. At the critical points, the work is checked – and signed off – by someone who can demonstrably judge whether it is really good and matches what was meant to be made.
- And when many teams work on it, up to the whole organisation?
- Shared ground rules for data protection and AI, training for everyone and clear agreements on how the teams work together. And room for every team to experiment, carried by a healthy culture of learning from mistakes.
From thirty years of practice
At IDEO I learned: numbers show what people do. Only observation and conversation show why they do it, what they wish for and what they fear. That is where my work begins. Some of what has come of it:
- Financial services
- A reimagined digital and physical banking experience at the ATM that remains unmatched to this day.
- Automotive supplier
- The future of electric mobility made tangible in a physical prototype with digital components, years ahead of the market.
- High-tech
- From gaming to home entertainment to ubiquitous computing – new forms of computer interaction through gesture control using visual 3D sensing (patented).
- Insurance
- Development of a global design system that lets a brand appear consistently worldwide. And its use in online insurance sales.
- Automotive, sales
- Concept and delivery of Europe’s first online shop in which manufacturer and dealers sell new cars together.
- Media
- Name, brand and first product design for a new TV provider, in three months. Today the market leader for IPTV in Germany.
- Software
- A trainable machine-learning platform that makes specialist texts usable at industrial scale.
- Automotive, machine learning
- Development of an embedded in-car entertainment prototype that predicts the behaviour of individual passengers from vehicle data and acts accordingly.
- Travel
- Implementation of a cloud-based digital asset management system and migration of several large image archives. Plus a highly complex taxonomy, roles and permissions, workflows and training.
Machines now generate texts, images and software. That turns efficiency into a commodity.
What makes the difference now is what you can sense in a piece of work: the effort, passion, character and intent of the people behind it – and the courage to pursue the unlikely.
Time to get things going.
After a first conversation, it’s clear whether there is a next step – and which one.
