Julie Vera's logo, an 8-bit, bright pink pixelated image of a coffee cup.
Home / Work

Work

Industry research organized by the decision it informed, not the deliverable it produced.

These are the shapes of the work. I'm happy to walk through specifics, methods, and findings directly. Send me a note →

AI-Generated Content & Synthetic Content Trust

When Do Users Trust AI-Generated Product Content?

AutoNation was evaluating whether AI-generated vehicle summaries could appear on a page where visitors are making a high-stakes purchase decision, and if so, under what conditions that content would read as trustworthy rather than suspect. I designed a mixed-methods concept test with 50 participants across desktop and mobile, using preference tasks and think-aloud protocols to understand not just whether people accepted AI-generated content, but why.

AutoNationAI ContentTrust Calibration
RoleAs principal investigator, I owned study design, stakeholder alignment across product and legal, synthesis, and executive communication of the findings.

0→1 Research Function Building

Building AutoNation's First Behavioral Persona Framework

AutoNation had no existing consumer research infrastructure, so before persona work could inform any product decision, I had to build the function that would produce it. I triangulated 30 in-depth interviews with a 500-participant survey to develop a behavior-first archetype system: one that describes not just who buyers are, but how they decide.

AutoNation0→1 ResearchOrganizational Vocabulary
RoleAs principal investigator, I led interview recruitment and fieldwork, designed the validating survey, and developed the archetype framework now used across product and design teams.

Archetype Deep Dive: The Skeptic

The Skeptic is one of six behavioral archetypes to emerge from the AutoNation persona framework: a trust-guarded, detail-oriented buyer who treats every claim on a vehicle listing as something to verify rather than believe.

AutoNationPersona
RoleI developed and validated this archetype within the broader framework, drawing on interview and survey data specific to trust-related purchase behavior.

Marketplace Search & Decision Support

Improving Search Relevance in a Two-Sided Marketplace

Monster's job seekers were losing confidence in search results, but the cause wasn't obvious from behavioral data alone. I ran a study evaluating perceived result quality directly, and traced the erosion of trust to specific weighting decisions in the ranking algorithm.

MonsterSearch & Ranking
OutcomeThe findings led to search-weighting changes that measurably improved perceived relevance across multiple markets.

When Faster Wasn't Better: Rethinking Swipe-to-Apply

Monster was considering a swipe-based application paradigm modeled on dating and shopping apps. I tested it against how job seekers actually think about applying, and found the mental model didn't transfer.

MonsterProduct Judgment
OutcomeThe research prevented investment in a product direction misaligned with user expectations, and surfaced unmet demand for interview-preparation tooling instead.

Monster Career Advice Microsite

The Monster career advice microsite had a rising bounce rate that the team assumed was a content problem. My research showed the cause was information architecture and visual design instead.

MonsterInformation Architecture
OutcomeThis finding prevented a costly content rewrite and redirected the team toward a design-led fix.

Internal Tools & Enterprise Research

UX Research Intern, Amazon (AWS), Summer 2022. This project examined internal sales tooling. The details remain internal to AWS and aren't published here, but I'm happy to discuss the shape of the work directly in conversation.