Go back

Designing trust in AI research workflows

Designed core workflows for collecting and organising large datasets, along with AI transparency.

Illustration showing two mocks up: one screen displaying the collector preview platform, another iphone mockup showing the artwork details.

Role

Product Designer — end-to-end UX/UI, prototyping, workflow design, and shipping decisions.

Scope

Research hub workflows, AI verification patterns, reporting templates, navigation, and reusable interface components.

Timeline

2023 – 2025

Summary

MosaiQ Labs is an AI-powered research platform used in high-stakes professional contexts. As the product evolved, one of the core UX challenges became trust: users needed to understand, inspect, and rely on AI-generated outputs quickly enough to use them in real decision-making.

I designed workflows and transparency patterns that made the system easier to navigate and its outputs easier to verify. This included improving how users worked with large datasets, how evidence was surfaced, and how research was transformed into repeatable deliverables.

What changed

Reworked the research hub structure to support larger datasets and faster navigation
Introduced AI verification patterns such as citations, source visibility, and clearer system states
Reduced friction through stronger grouping, tabs, filters, and repeatable reporting modules

Outcome

The result was a more usable and trustworthy research environment: users could orient themselves more quickly, inspect outputs more confidently, and move more smoothly from research to report creation.

Reflection

This project reinforced that in AI products, trust is not a layer added at the end — it is part of the interaction model itself. Good design helps people understand what the system is doing, what it knows, and when to rely on it.

Key screen

Research hub refresh

I reworked the project page to support larger datasets and faster navigation. The redesign introduced clearer grouping, tabs, and filters so users could orient themselves more quickly and move between sources without losing context.

Key screen

Citations and verification patterns

To make AI-generated outputs more trustworthy, I designed citation and evidence patterns that let users inspect claims, understand where information came from, and verify outputs without leaving the workflow.

Key screen

Reporting template

I introduced reusable templates to help teams convert research into consistent outputs. This reduced repetition and supported a smoother research-to-report workflow.

Go to the full case study

Back to the top

Next Project