AI-ENABLED SERVICES · SERVICE DESIGN · 2026
What do you design when the interface starts to disappear?
AI is moving deeper into services — analysing information, recommending actions and influencing decisions long before someone reaches a screen.
I worked on a research-led service design project exploring what that shift means for the people designing and delivering AI-enabled services: where human judgement still matters, how responsibility changes, and how those decisions can become visible enough to design deliberately.
Service design
Research synthesis
Framework Design
Facilitation tools
Experience standards
01 / The shift
The interface was no longer the whole experience.
Traditional UX gives us visible moments to work with: a screen, a form, a journey, an interaction. But when AI begins operating inside a service, consequential decisions start happening elsewhere — in how a signal is interpreted, whether an output is trusted, what gets prioritised, and when someone intervenes.
The person using the service may never see those moments. They still experience their consequences.
AI governance was also an experience-design problem.
02 / RESEARCH & SYNTHESIS
We followed the decisions behind the experience.
Instead of starting from individual interfaces, we looked upstream — at emerging behaviours, workflows and the moments where AI was beginning to change how decisions were made.
The challenge wasn’t a lack of information. It was connecting different kinds of evidence without flattening them into a list of findings.
The design question became less about what AI could do — and more about where it should stop doing it alone.
I worked across signals, behavioural patterns, decision moments and experience risks to trace how one observation could travel through a system and eventually affect someone’s experience.
03 / FROM FRAMEWORK TO PRACTICE
Not every decision should move at the same speed toward automation.
We needed a shared way to talk about how much responsibility AI should carry in different moments of a service.
Rather than treating greater autonomy as automatic progress, we explored it as a spectrum — from AI supporting human judgement to systems acting with increasing independence.
The useful question wasn’t “How autonomous can this become?”
It was:
“At what point does assistance become authority?”
A principle only matters if someone can use it on a Tuesday afternoon.
The framework was translated into practical guidance and facilitation tools: ways for teams to ask what role AI was playing, who could challenge it, what happened when it was wrong, and whether a meaningful human path still remained.
04 / WHAT I TOOK FROM IT
The interface wasn’t disappearing. It was expanding.
This project changed how I think about AI in service design.
When technology begins making decisions inside a system, designing the experience can’t stop at the moment someone touches an interface.
It also means designing what happens around it:
what the system can infer, what a person can question, where judgement sits, and how responsibility remains visible.
Sometimes the most important part of an interface is the decision that happened before it appeared.