Working with Generative AI

Thoughtful. Rigorous. Ethical.

In an era where AI is reshaping how work gets done, I use it deliberately. To handle the mechanical load of research so I can invest more deeply in the thinking, and as a consistency check on my own judgement. Human cognition comes with shortcuts; AI, applied thoughtfully, can help surface patterns and potentially reduce certain cognitive biases that we as humans inevitably carry.

It won't eliminate bias — nothing does — but that's precisely why I treat every AI application as a methodological choice, not a convenience: does it preserve the integrity of the insight, respect participant privacy, and produce something I can stand behind?

This portfolio is itself a small example of that philosophy, built with AI assistance, but the research, the insights, the relationships, and the impact behind it are entirely my own.

Methodological rigour
Every AI tool is evaluated against research standards before adoption
Participant privacy
Data handling and consent considerations precede any AI application
Accountability
Interpretation, nuance, and professional judgement remain mine
How I use AI in my work
Applied thoughtfully, across the research process
01
Live interpretation & transcription

At OKX, to accelerate research delivery and minimise cost, I trained local managers to conduct interviews and usability tests in their native languages. To maintain research quality while working remotely, I used live AI interpretation to follow sessions in real time — understanding what was being asked, how participants responded, and where to probe further. The AI-generated transcript and translation also supported my post-session debrief meetings with local managers, helping me seek clarifications and surface follow-up questions I would otherwise have missed.

02
AI translation of multilingual survey responses

At OKX, I used AI translation to process qualitative survey responses received in 5 non-English languages across a 10,000-response survey, dramatically reducing turnaround time and translation costs while maintaining analytical quality. All translated responses were manually reviewed and coded before analysis.

03
Structured interviews run by AI

An AI-powered interviewer designed to conduct structured UX research interviews at scale — maintaining consistent question framing, probing on key themes, and capturing responses for analysis. Built to complement, not replace, in-depth human-led sessions where depth and nuance matter most.

In development
What's next
Coming soon
Thematic analysis to academic standards

AI-assisted thematic coding grounded in academic frameworks, ensuring analytical rigour at scale without sacrificing interpretive depth.

Coming soon
Diary study data capture tool

A purpose-built tool for capturing longitudinal diary study entries, reducing participant friction and improving data quality over time.