Defining AI strategy for home search
When competitors started launching AI chat features, Realtor.com wanted a fast response. The brief was simple: build a chatbot. I led the strategy work that reframed that request, shaping three concept directions that became RealAdvice, an assistive experience positioning AI as an empathetic guide rather than a gimmick. Working across product, engineering, research, and data science, I aligned leadership on what "assistive AI" meant at Realtor.com and created the foundation for future work across search and discovery.
Company
Realtor.com
Year
2024
Role
Lead Product Designer
Timeline
4 months
01 CHALLENGE
From chatbot mandate to strategic opportunity
At the time, competitors were announcing AI features one after another. The ask was straightforward: "We need a chatbot like that."
But buying a home isn't a chat problem. It's a confidence problem. Buyers were navigating a volatile market, tight budgets, and high stakes. They worried about overpaying, missing hidden gems, or choosing the wrong area for their lifestyle. A chatbot that answered listing questions wouldn't fix that.
My first exploration reflected that thinking. I mocked up a simple chat assistant to answer listing questions and suggest nearby options. In testing, it felt transactional—more like a faster filter than a smarter guide. It solved for curiosity, not confidence.
That exploration clarified something deeper. The opportunity wasn't making AI chat better. It was making home search feel more human.
Key questions I focused on
Would a chatbot actually address real buyer needs, or just match competitor features?
How could AI support decision-making instead of adding noise?
What level of transparency and control would people need to trust AI in this context?
What would make our approach meaningfully different, not just "we have AI too"?
02 RESEARCH
Understanding buyer anxiety
I started by reframing the problem: what was really getting in the way of people finding a home. It wasn't access to listings. It was confidence.
In research and previous discovery work, the same patterns kept showing up:
Buyers were anxious about affordability and long-term risk
They were afraid of missing out on the "right" home
They felt unsure how to weigh trade-offs between price, commute, schools, space, and lifestyle
People were bouncing between apps, messaging agents, and driving neighborhoods—still unsure if they were missing something
To bring the team onto the same page, I mapped the buyer journey in Miro and layered in where AI could help, from early exploration to making decisions. The board captured the core insight: affordability anxiety, fear of missing out, and the need for guidance over automation.
Working with engineering and data science, I looked at what was realistic. We could connect structured listing data, filters, and preferences, but not judgment or nuance. That limitation became the opportunity. Instead of making AI predict everything, I designed it to guide.
03 SOLUTION
Turning insights into concepts
I led the design strategy that turned these insights into tangible product concepts. Together with the team, I built three directions. Each addressed a different buyer need and showed how AI could assist rather than automate.
Concept 1 — Integrated in Search
I explored embedding AI directly in the search experience to surface hidden gems and highlight neighborhood trade-offs buyers often overlook. By connecting structured data with behavioral intent, it helped users discover "total matches" that balanced budget, lifestyle, and proximity.
User value: Uncover overlooked listings, reduce fear of missing out, and feel confident they're seeing the full picture.
Concept 2 — Lifestyle Compatibility Profiler
This concept explored how AI could help buyers understand how well a home or area fit their lifestyle. The profiler balanced "nice-to-haves" against affordability, translating vague wants into clear, actionable recommendations.
User value: Make realistic trade-offs, adjust expectations, and discover homes that align with personal priorities.
Concept 3 — Building User Preferences
I designed this concept to show how the experience could evolve over time. As buyers refined what mattered most, AI would adjust in kind, suggesting comparable homes, better-fit neighborhoods, or new financing paths.
User value: Personalization that grows with you, helping buyers move from exploration to confident decisions.
RealAdvice — The story that tied it together
To help leadership and teams visualize this future, I created a narrative around a first-time buyer named Sophie and her partner Kyle. The story showed how AI could guide them from early anxiety to clarity, helping them compare homes, document visits, and weigh trade-offs with confidence.
Instead of "Here's what I found," RealAdvice says "Here's why this fits your goals." That shift from answers to understanding became the north star for how we talked about AI in the product.
04 impact
From vision to direction
The work didn't end with concepts. It built alignment. I presented the RealAdvice vision and the three concept directions to product, engineering, and design leadership. The conversation shifted from "what can we build fast?" to "what do we want AI to actually do for people buying a home?"
That shift gave teams a shared language for assistive AI at Realtor.com. It connected research, data science, and design into one narrative instead of separate experiments.
We didn't launch a chatbot. We launched a shared understanding of how AI should serve people—one that shaped Filter Foundations, natural-language search, and every AI initiative that followed.
05 REFLECTION
Strategy before speed
Biggest lesson: the most valuable work happens before pixels ship. By reframing the problem from "what chatbot should we build?" to "how can AI help buyers feel confident in their decisions?", I turned urgency into clarity. Making the strategy tangible through concept prototypes and storytelling, not just slides, helped leadership see a future they could act on. Those artifacts became the blueprint for the 2025 Filter Foundation and AI Search sprints, proof that empathy and clarity can drive scale.
The RealAdvice vision showed that strategy isn't about predicting the future. It's about giving teams something real to build toward. The best responses to competition don't match features. They build confidence and trust where it matters most.