Scaling Our Impact with AI: Research Planning Co-Thinkers
Focus
Internal UX Research Enablement
AI Integration
Client: Fortune 500 retailer
Role: Strategist, Educator, and Builder
Impact: Scaled research consultations, improved planning quality, and democratized access to evidence-based practices
With AI-powered co-thinkers, we gave teams faster, smarter ways to plan studies and sparked a new way of working: one that saves time, builds confidence, and keeps humans where they matter most.
The Challenge
My Role
Our centralized UXR team of 4 supports a growing community of 40+ researchers across the org.
Office hours were a go-to for coaching — helping teams clarify questions, tighten goals, and think out loud. But demand outpaced our capacity, and those high-value dialogues didn’t scale.
We needed a way to deliver the thinking and structure of office hours — without the human bottleneck.
Designed and led the AI-driven research enablement strategy
Created a custom GPT tool rooted in team standards and best practices
Expanded the assistant into a living system with real-time academic updates
Scaled impact by turning consultations into reusable co-thinking frameworks
Enabled 40+ researchers to plan better, faster, and with greater confidence
Phase 1: Building the Research Planning Assistant GPT
I created a custom GPT trained on our internal research standards and templates.
Its purpose?
Simulate the exact conversation we’d have during an intake session and output a first-draft research plan — structured, aligned, and ready for review in office hours.
✅ Benefits:
Reduced prep time for teams
Higher-quality research plans
Clearer alignment with internal best practices
Freed our human team for more strategic engagements
Phase 2:
Expanding into AI-Human Interaction
Inspired by Stanford’s “UI/UX Design for AI Products” course and growing internal demand for AI-infused tools, I applied the same co-thinker concept to AI research itself.
I built a second GPT focused on Human-AI interaction — but didn’t stop there.
To keep it current, I:
Wrote a custom Python script
Scraped an open academic repository for the latest HAI research
Set up monthly updates to feed the GPT’s knowledge base
Result: A thinking partner that draws from cutting-edge research to guide AI feature design and exploration — without requiring everyone to read a dozen papers a month.
Outcomes
Scaled consultation: More people got the help they needed — when they needed it
Increased research maturity: First drafts came in stronger, faster
Democratized best practice: Tools were accessible across levels and teams
Freed up team capacity: We focused where the human layer mattered most
Next Steps
We’re extending the co-thinker model to new frontiers:
First-Pass Critique Framework
We’re building a feedback engine to give early-stage concept critiques — helping teams catch low-hanging UX issues before involving real participants. It’s not a replacement for users, but a smart filter to refine what should be tested.Simulated Omnichannel Environments
We’re experimenting with AI-driven simulations to prototype and evaluate complex, cross-channel experiences. Think call centers, web, mobile, and retail interactions — stitched together into a researchable environment before full deployment.