Designing warmth into the age of AI

I build AI systems that keep people at their centre. Between technology and empathy.

Hi, I'm Wout. I work in the space between AI and people.

I am a designer specialised in designing use cases for AI systems themselves, from multi-agent architectures and vector databases to conversational avatars. But the building is never the point. What interests me is the moment technology meets someone’s life. A person with dementia starting a music session on their own. A caregiver hearing a life story they never knew was there.

I call this position the Technology Orchestrator. I am technically fluent enough to build complex systems while remaining human-centred enough to ask what they are for. Unlike designers who engage with technology primarily at the interface level, I work inside the technological system itself, shaping use cases, data flows, agent interactions, and experiences simultaneously. This allows me to influence technological decisions before they become fixed constraints for users.

My main developed expertise areas are Technology & Realization and Business & Entrepreneurship. Together they allow me to move between creating technology and creating the conditions for it to succeed. I enjoy identifying opportunities, reframing challenges, and translating research into solutions that personal healthcare can actually implement, sustain, and create value from.

This position emerged throughout my development. I came to TU/e from creative technology, a maker and engineer at the core, focused on digital technologies. Aqua-morph taught me that research and design are about creating questions and arguing for potential impact. The helmet project pulled me far outside my technical comfort zone into behaviour, participation, and societal change. Het Ritme brought both sides together through warm technology and dementia care, and that is where I found my place.

Three beliefs run through everything I make. It must be functional: it has to work, technically and in practice. It must be achievable: grounded in what can realistically be built and adopted. And it must be meaningful: creating genuine value for the people it is designed for and with. Care technology is where these three principles are constantly in tension and none of them can be sacrificed. That is exactly why I work there.

In teams I often take the bridging role, translating between specialists who do not share a language and stepping forward when direction is unclear. I have learned to voice my ideas instead of holding them back to preserve harmony, a habit I first identified in my master reflections and have deliberately worked on ever since.

My aesthetic practice has developed through movement and sound. Aqua-morph built a vocabulary of bio-inspired motion as design expression. Sound of Smart Things showed me how sonic rhythm and timing carry meaning without visual cues. OpMaatje applied both: the warmth of a conversation comes as much from voice pacing as from the avatar’s visual design, and the era-styled living rooms were chosen specifically to trigger reminiscence through atmosphere. Visual form-giving is where I continue to grow deliberately. Throughout my studies I noticed that I could articulate technical architectures and interaction concepts more readily than communicate them through visual form. To address this I rely on prototyping, storyboarding, and iterative testing to externalise ideas early, and the design of this portfolio is the most considered visual work I have produced.

My approach to design is strongly shaped by Research through Design. I build working prototypes early and use them as research probes, not to validate concepts but to discover things that could not have emerged from interviews or assumptions alone. Prototypes are instruments for generating knowledge: by putting artefacts into the world, I uncover insights that shape both the design and the research. The recurring tension is the pull toward deep technical exploration. I learned to keep it purposeful by treating those investigations as deliberate probes with a clear design question at their centre.

Wout Vissers presenting on stage at the Warm Technology conference

Warmth is not a design aesthetic

Every major technological shift follows the same pattern. Big promises of efficiency arrive first, and the human implications are addressed later. The current development of AI is repeating that pattern. The stakes feel higher this time because AI does not simply automate tasks; it increasingly shapes how people are seen, heard, remembered, and cared for.

Engineering culture often defaults to efficiency as its primary metric. Yet there is rarely a shared language for the human dimension of a system, and rarely someone responsible for defending it. I believe designers must take that role. Not only by evaluating technologies after they are built, but by actively shaping how they are built in the first place. Decisions about memory, agency, autonomy, and personalisation are design decisions as much as technical ones.

Rens Brankaert's distinction between cold and warm technology names this challenge precisely. Warm technology focuses on what people can still do, supports them emotionally, and feels genuinely personal. Rather than replacing capabilities, it strengthens them.

My contribution to this discussion lies in the technologies themselves. Many designers advocate for warmth from the outside; I want to help shape it from within the systems that increasingly determine people's experiences. As AI becomes embedded in healthcare, designers need to participate in conversations about architectures, data, and intelligent behaviour, not only interfaces. If designers are absent from those decisions, systems will optimise for efficiency long before they optimise for human experience.

That is not an abstract commitment for me. During an OpMaatje session in a care home in Waalre, I watched a caregiver discover a life story they had never known about a resident they had supported for months. The AI created the conditions for a deeper relation within a care setting. That is what it looks like to design warmth from inside the system.

Care is one of the fastest digitalising sectors in Europe. Caregiver shortages and ageing populations make new technologies inevitable. Many emerging solutions are technically impressive but humanly thin. AI creates an opportunity to personalise care at an unprecedented scale, but only if designers actively guide how that personalisation is realised. The question is no longer whether AI will enter care, but what kind of care it will create.

Warm technology is not about maximising comfort. In care, the warmest solution is often not the most efficient one: a conversation takes longer than an automated reminder yet creates value that efficiency metrics cannot capture. Designing for warmth means holding organisational pressure and human need in balance, not choosing between them.

That balance is where I aim to position myself. I believe the next generation of care technology will not be defined by automation alone, but by how well it augments human capabilities and relationships. My role as a designer is to help ensure that the systems shaping people's lives remain not only intelligent, but genuinely human.

  • Technically functional, humanly empty
  • Built for throughput
  • Treats people as data points
  • Optimises what can be measured

Toggle to feel the difference.

Warm AI does not simulate empathy. It is designed so empathy has room to exist.

Winning the Warm Technology Prize is an early signal I’m on the right path.

A moment of your own

OpMaatje is a conversational AI avatar for people with dementia in residential care homes. It uses music, reminiscence, and personalised conversation to surface autobiographical memories and strengthen identity, at a time when caregivers are under increasing pressure.

One of two avatars, Daan or Eva, leads the conversation. A Together mode lets a caregiver or family member join, sharing what the avatar has learned about that person. Music opens the session; personal memories are the ongoing thread.

Daan, Eva and Together Mode

Two distinct avatar personalities, plus a mode where a caregiver/familly member can join the conversation. You choose who feels right.

Multi-agent architecture

Specialised agents coordinate the conversation, retrieve personal memories, act upon those memories, and keep the dialogue safe.

Music as the entry point

A familiar song opens the door. Personal memories become the ongoing thread of every conversation.

Daan

OpMaatje avatar

Demo conversation

Het Dorp, Wim Sonneveld

Interactive demo

A song opens the door. A memory walks through it.

This conversation shows the shape of a real OpMaatje conversation. The avatar starts from a song the person with dementia remembers, then follows the memory wherever it leads, gently and at your own pace.

MDCMulti-agent architecture, use of multiple AI models, functioning back and frontend
U&SCo-design with residents, context mapping, ethical research with vulnerable populations
T&RFunctional deployed system, field-tested in a care home
B&EDesigned for caregivers and institutions, not just residents
C&AEra-styled avatar environments, non-uncanny character design, voice pacing and music as emotional entry point

Built with JavaScript and React, multi-agent coordination. Tested in a real care home.

Warm Technology Challenge 2026, jury prize

The full OpMaatje story

Five areas. One integrated practice.

Every project, placed by the expertise it drew on. The closer to the centre, the more areas it integrates.

The combination that has defined my work most directly is T&R, B&E, and U&S together. Knowing how to build a multi-agent memory architecture and how to run co-design session with people with dementia are skills that rarely come from the same person. In care technology, both are necessary at once. Adding an entreperneurial attituate, and making descisions values based you get an designer that focuses on the technollogy, users and the entire stakeholder landscape.

A pentagon with the five expertise areas at its corners. Each project and course is placed by the mix of areas it drew on; works near the centre integrate more areas.

  • OpMaatje (FMP): Business and Entrepreneurship level 5 of 5, Technology and Realization level 5 of 5, Math, Data and Computing level 4 of 5, User and Society level 4 of 5, Creativity and Aesthetics level 2 of 5
  • Het Ritme (M2.1): Technology and Realization level 4 of 5, Math, Data and Computing level 4 of 5, Business and Entrepreneurship level 3 of 5, User and Society level 3 of 5, Creativity and Aesthetics level 2 of 5
  • Let's Talk About Helmets?! (M1.2): User and Society level 5 of 5, Business and Entrepreneurship level 4 of 5, Creativity and Aesthetics level 3 of 5, Technology and Realization level 1 of 5, Math, Data and Computing level 1 of 5
  • Aqua-morph (M1.1): Technology and Realization level 5 of 5, Creativity and Aesthetics level 4 of 5, Business and Entrepreneurship level 1 of 5, Math, Data and Computing level 1 of 5, User and Society level 1 of 5
  • refrAImed (M1.2): User and Society level 4 of 5, Creativity and Aesthetics level 3 of 5, Technology and Realization level 2 of 5, Business and Entrepreneurship level 1 of 5, Math, Data and Computing level 1 of 5
  • Design Entrepreneurship (Course): Business and Entrepreneurship level 5 of 5, Creativity and Aesthetics level 2 of 5, User and Society level 2 of 5, Math, Data and Computing level 1 of 5
  • The Sound of Smart Things (Course): Creativity and Aesthetics level 5 of 5, Technology and Realization level 2 of 5, User and Society level 2 of 5, Math, Data and Computing level 1 of 5
  • Digital Twins (Course): Technology and Realization level 4 of 5, Math, Data and Computing level 3 of 5, User and Society level 2 of 5, Business and Entrepreneurship level 1 of 5, Creativity and Aesthetics level 1 of 5
  • Designing with Advanced AI (Course): Technology and Realization level 4 of 5, Math, Data and Computing level 4 of 5, User and Society level 3 of 5, Business and Entrepreneurship level 2 of 5, Creativity and Aesthetics level 1 of 5
  • Embodying Intelligent Behavior (Course): Math, Data and Computing level 5 of 5, Technology and Realization level 3 of 5, User and Society level 2 of 5, Creativity and Aesthetics level 1 of 5
  • Values-Based Leadership (Course): Business and Entrepreneurship level 5 of 5, User and Society level 3 of 5, Creativity and Aesthetics level 1 of 5
B&EBusiness & EntrepreneurshipT&RTechnology & RealizationMDCMath, Data & ComputingC&ACreativity & AestheticsU&SUser & Society
B&Emain area

Business & Entrepreneurship

I think B&E is what forces my work to actually land somewhere. The helmet project taught me that understanding who can really move the needle, schools, campaign organisations, national safety bodies, is what makes an outcome adoptable instead of just interesting. Design Entrepreneurship built a habit of asking whether the problem is even the right problem before touching a solution. And with Het Ritme, I think that same habit is what eventually grew into OpMaatje: two care partnerships, a jury prize, and a real path toward deployment. That does not happen without constantly asking what actually makes sense in the world outside the project.

  • PESTEL + stakeholder mapping
  • Aumens startup collaboration
  • De Wever + De Meerpaal care partnerships
  • Design Entrepreneurship client sprint
T&Rmain area

Technology & Realization

For me, building something is how I understand it. With Aqua-morph I had to design the fabrication method before I could design the artefact, which I think says a lot about how I work. In Digital Twins I built a sensor node from scratch: hardware, embedded code, environmental sensing, and an SVM classifier trained on real data. With Het Ritme and OpMaatje I applied the same thinking to software, agentic conversation flows and a backend I maintained myself, until the system was stable enough to run sessions with people with dementia without anyone watching over it. The prototype is how I think; the field test is how I know if I was right.

  • Hydrogel actuator fabrication method
  • Custom embedded sensor node
  • SVM walkability classifier
  • Field-tested agentic system
MDC

Math, Data & Computing

I think one of the most important things this area taught me is that AI should not be a black box. Building the full pipeline for Daisy, from data collection and pose estimation to a trained classifier and real-time feedback, meant I understood every decision I made and why. ConvoKeeper added vector embeddings and retrieval logic for matching care residents to topics and activities. OpMaatje built further with multi-agent coordination and memory systems across multiple models. Being able to open the hood and explain what is happening inside a system changes how I design with it.

  • End-to-end ML pipeline (Daisy)
  • Vector embedding architecture
  • Multi-agent orchestration (OpMaatje)
U&S

User & Society

I think U&S has pushed me furthest outside of what I was comfortable with. Researching helmet use among teenagers meant cold-calling almost every school in Eindhoven and learning that behaviour change has almost nothing to do with product design. Working with people with dementia required something else entirely: letting go of strict protocols, reading the mood of the session, and accepting that a relaxed conversation sometimes mattered more than a research question. Het Ritme was tested with ten participants, OpMaatje with residents across three care organisations. Each group reminded me that there are no user types, only people with their own histories.

  • Ethics with minors and vulnerable adults
  • Co-design workshops across age groups
  • Dementia user tests and research
C&A

Creativity & Aesthetics

I think my aesthetic sense lives more in movement and sound than in visual form. Aqua-morph was really about what motion can communicate, building a vocabulary of eight bio-inspired types and arguing for what each one expresses. Sound of Smart Things showed me that rhythm and timing carry meaning visuals simply cannot. In OpMaatje those two things came together: the warmth of a conversation has as much to do with voice pacing as with how the avatar looks. This portfolio is probably the most considered visual work I have made.

  • Aqua-morph motion vocabulary
  • Sound of Smart Things sonic identity
  • OpMaatje avatar + era environments

Where I’m coming from

My FMP is not the beginning of this work. Each project added expertise and knowlege building towards my FMP

  1. M1.1

    Aqua-morph

    Shape-changing hydrogel actuators. A study in the aesthetics of movement.

    C&AT&R
  2. Early M1.2

    refrAImed

    Grandchildren and grandparents, connected through photos turned into digital art.

    U&SC&AT&R
  3. M1.2

    Wear a Helmet

    Helmet use among teens on fatbikes. The research led to a workshop, not a product.

    U&SB&E
  4. Prep FMP

    Het Ritme

    A conversational avatar tested with ten people with dementia. The starting point of my FMP.

    MDCU&ST&R
  5. FMP

    OpMaatje

    A warm avatar for music reminiscence. Jury prize winner, now continuing toward real care homes.

    B&ET&RMDCU&S

Alongside the projects, courses build breadth across the expertise areas. The courses mostly had a focus on T&R and B&E, most connected to the AI/smart systems domain.

Where I’m going

I am building toward a career in care technology, where the gap between what AI can do and what people need from it is most visible. Digitisation in care is inevitable; my aim is to make it warm. OpMaatje showed me where I belong as a designer, and the next step is taking that further at Aumens, where care technology meets the market. Making warm technology the standard, not the exception.