Understanding & expectations
Do people understand what the service can do, what information it uses and where its limits are? Observe how those expectations affect a real task.
UX research for AI services
Help people use AI with understanding and control.
Research how people experience an AI-assisted service, from their first expectation to the moment they need help.
FabUX brings behavioural research and usability testing to AI-assisted public services, customer journeys and workplace tools. We help teams find where an experience supports people and where it creates confusion, misplaced confidence or extra work.
Work can begin in discovery, test a prototype or examine a live service. We can work directly with your team or deliver focused research within a larger programme.
Research questions
What happens when people rely on the service?
Trust & uncertainty
Do people question an answer when they should? Test how explanations, sources and uncertainty are interpreted, including when a confident answer is wrong.
Control & human support
Can people review, correct or undo an action, and reach a person when needed? Examine the handover as part of the whole service journey.
Inclusion & recovery
Research with people who have different access needs, language preferences and digital confidence. Explore what happens when the service misunderstands them or cannot help.
Our approach
Watch the task, then revisit the decisions.
We begin with realistic tasks and natural behaviour, then use retrospective questions to explore what people understood, expected and relied on. Remote and face-to-face research can be blended around the participants and the service.
We agree suitable scenarios, participation and recording arrangements, using test information where personal or sensitive material is unnecessary. Prototype behaviour and simulated responses are clearly distinguished from a live system.
Findings connect observed difficulties with practical changes to content, interaction, support and service design. We make the limits of the sample and evidence explicit.
This is research into the human experience. Technical model evaluation, security assurance and formal compliance assessment require their own specialist work.
Foundation
Practical guidance alongside the research.
Foundation is our collection of AI governance resources, including a free AI Use Policy and guidance on implementation, human review and privacy. It helps teams turn their intentions into clear working practices.
Explore Foundation and the free AI Use Policy
For how your organisation is understood by answer engines and AI agents, our Content & Answer Engine Reviews provide a separate, complementary service.
Related services
Continue the investigation.
Connect AI research with the wider service, its content and the people using it.
- User Research & Usability Testing
Understand people’s needs, test real tasks and find where content, access or confidence breaks down.
- Public Services
Behavioural evidence for discovery, service improvement and inclusive public services, from first contact to getting help.
- Content & Answer Engine Reviews
Make your organisation easier for people and answer engines to understand, trust and represent accurately.
Where does AI enter your service?
Tell us what people need to do, what you are testing and the decisions your team needs evidence for.