Accessible only to conference ticket holders.
Log in Create account
For 90 days after a conference, only paid ticket holders can watch conference videos. After that, all Gold members have access.
If you have purchased recording access and cannot see the video, please contact support.
Conducting pre-research with AI agent personas: Pressure-testing concepts for expert workflows
This video is featured in the AI and UX skill growth playlist and 1 more.
Summary
Designers in complex domains like finance, healthcare, or government are often asked to design for expert workflows they barely understand, while their subject-matter experts are too busy to join every iteration. This case study shows how niche AI “agent personas” became an always-available, domain-aware first pass on ideas—helping a UX professional in fintech to pressure-test concepts, understand upstream/downstream impacts, and arrive at better prototypes faster. In this work, AI serves explicitly as a pre-research layer, not a substitute for real users, formal research, or usability testing, ensuring that human insight remains at the center of the design process.
Key Insights
-
•
AI personas can help UX teams pre-emptively pressure test complex workflows before SME access.
-
•
Structuring AI personas with detailed role, goals, pain points, and boundaries ensures specific and relevant feedback.
-
•
Including a confidence summary in AI responses helps users distinguish between grounded facts and inferences, mitigating hallucinations.
-
•
AI critiques of wireframes expose operational friction and workflow gaps, not just surface usability issues.
-
•
Integrating AI with codebases improves understanding of data lineage, user entitlements, and ecosystem dependencies.
-
•
AI cannot guarantee policy compliance or capture all organization-specific exceptions.
-
•
Early AI involvement fosters confident preparation and sharper questions in SME reviews.
-
•
AI-driven expanded scope can cause scope creep, needing prioritization and feasibility discussions with technology teams.
-
•
AI responses are generally repeatable and consistent when grounded in stable persona definitions.
-
•
AI pre-research helps challenge operational workflows, uncovering inefficiencies that improve overall UX beyond UI enhancements.
Notable Quotes
"Can AI help us do better in pre-research for expert-heavy workflows before we actually get access to the subject matter experts?"
"The real problem is that the team is already making decisions in incomplete conditions."
"Generic advice is usually too broad for a complex workflow such as escalation or compliance."
"AI personas can be valuable as bounded pre-research tool helping us think earlier and reduce avoidable gaps."
"The confidence summary was added to remind us that the AI response is provisional and needs verification."
"When AI got it wrong, it did so in ways that matter operationally, skipping required compliance or assuming outdated workflows."
"The biggest improvement was confidence — walking into a room with senior stakeholders prepared to ask sharper questions."
"AI cannot replace user research, usability testing, or domain expert validation."
"AI helped me challenge operational workflows and ask why approvals happen multiple times across systems."
"Running the same prompt with a defined persona produces consistent and reliable AI responses."
Or choose a question:
More Videos
"Android’s accessibility tools started later than Apple’s, so the community is smaller and less documented."
Sam ProulxUnderstanding Screen Readers on Mobile: How And Why to Learn from Native Users
March 27, 2023
"The seventh generation principle calls for long-term decision making that transcends self-interest."
Jayne Engle Tanya Chung-Tiam-FookCivic Design for the Next Seven Generations—A Discussion on Sacred Civics
August 25, 2022
"Visual design quality actually affects perceived trustworthiness."
Daniel J. RosenbergDesigning with and for Artificial Intelligence
August 11, 2022
"Multi-stage rockets are like the double diamond; you jettison valuable context at every stage and lose signal."
Erika FlowersThe Handoff is Dead: Design-Led Engineering with AI Agents
March 4, 2026
"It’s not enough just sharing customer stories and insights; we need to get into the weeds and collaborate with product and tech partners."
Liwei DaiThe Heart and Brain of the AI Research
March 31, 2020
"Nobody’s going to use digital calendars because they’re just more work and no benefit."
Sam LadnerData Exhaust and Personal Data: Learning from Consumer Products to Enhance Enterprise UX
June 8, 2016
"There’s been a huge convergence because everybody’s got a mobile device now and expects a certain kind of experience."
Amy MarquezINVEST: Discussion
June 15, 2018
"It’s breath not death. Don’t try to solve for the system. Try to understand the request and help the business."
Jacqui FreyFlow and Superfluidity for Design Orgs
November 7, 2018
"They walked. The locals were most likely right."
Brendan JarvisFraming Tomorrow by Questioning Today
June 8, 2022
