An AI Perspective · Tristan Machado

AI & user research

How do you see AI shaping user research?

I see AI as a research assistant, not a researcher. It’s great at the heavy lifting that slows studies down: transcribing sessions, organizing notes, clustering early themes across many interviews, and turning findings into prototypes stakeholders can react to, as I did with Figma Make.

That frees up more of my time for the parts of research that matter most: building rapport with participants, asking the follow-up question, and noticing what people don’t say. It’s in those moments that I uncover the nuance behind a user’s pain point, like the context that created it and the emotions tied to it, which AI can easily miss.

AI can speed up synthesis, but empathy and judgment are still the researcher’s job.

I treat every AI-suggested theme as a hypothesis, not a finding, and go back to the raw data to confirm it. AI can summarize what was said, but it can’t read a participant’s hesitation in the moment. And because research often involves sensitive information, participant consent and privacy come first: recordings and transcripts only go into tools that are approved to handle that data.

AI tools & design

What are your views on modern AI tools in design?

I see modern AI as a powerful design accelerant. It speeds ideation, surfaces patterns from research, and automates repetitive craft tasks so designers can focus on strategy and human insight.

Tools like generative interfaces and assisted prototyping (e.g., Figma Make) let us explore many directions quickly, validate concepts faster, and produce higher-quality artifacts for stakeholder alignment.

AI is best used as an augmentation, not a replacement, of domain expertise.

Human judgment, ethical considerations and contextual knowledge remain essential to vet outputs and ensure they serve real people.

AI & the future of UX

How do you see AI shaping the future of UX?

For systems designers, AI will reframe the discipline from crafting individual screens to orchestrating resilient experience architectures: we’ll define intent, data contracts, interoperability rules and experience policies that govern how models behave across workflows.

AI will automate variant generation, localization and routine decisioning, enabling systems to surface context-aware guidance (e.g., role, tenure, risk signals) while enforcing compliance and auditability.

UX research within systems work will expand into model auditing, bias testing and telemetry design, ensuring automated behaviors are measurable, explainable and composable into larger enterprise flows.

AI, engineers & me

How would you work with engineers in an AI-forward environment?

I’d treat engineers as co-pilots in a shared experimentation lab: we’d agree on success metrics upfront, instrument everything, and iterate in short, measurable loops. Practically, that looks like pairing on data needs and synthetic test cases, building feature flags and human-in-the-loop checkpoints, while automating rollback criteria for model drift or harmful outputs.

I push for operational rigor (model versioning, clear ownership of training data and lightweight governance docs) so design hypotheses become reliable product behavior.

Collaborate early, measure continuously, and design for graceful degradation.

That way, the product remains useful and trustworthy when AI is involved.

See it in practice: in the Membership Rewards case study, I used Figma Make to turn confidential research into a shareable, testable prototype, and in Thoroughbase I paired an AI-generated prototype with a hands-on heuristic review.