Case study 01 — AI / LLM

AskZ: designing Zillow's first LLM product

A conversational beta for the highest-stakes consumer decision there is, and the foundation for Zillow's AI Search today.

RoleDesign lead (and only designer)
Team3 PMs · 3 eng teams · researcher · content strategist
OutcomeGreenlit Zillow's AI copilot; foundation for AI Search

Context

In 2023, Zillow decided to find out what LLMs could do for the hardest part of home buying: the money. As a senior designer on the Dream and Search team, I was asked to lead design for AskZ, a finance-focused chatbot to help people with financial questions on the way to a home, and the company's first LLM product. I was the design lead and the only designer, partnering with three PMs, three engineering teams, a researcher, and a content strategist, with AI, Search, Mortgage, and Legal leadership all at the table. It shipped on web and mobile web, with webview integrations for iOS and Android.

Real estate is an unforgiving domain for generative AI. A hallucinated restaurant recommendation is a bad dinner; a hallucinated answer about mortgages or property taxes can distort a six-figure decision. The quality bar had to be set accordingly, and leadership set it high.

How I worked

This was a new technology for the whole organization, so I embedded directly with engineering rather than working downstream of them, following their investigation of building on our own LLM versus a GPT integration, and what each meant for guardrails. I needed to understand what the model could actually do, because those properties are the design material. I also built empathy the unglamorous way: by becoming a heavy LLM user myself (for work, for personal questions, even recipes) until I had instincts for where these systems delight and where they quietly fail.

Before any interface work, I ran a collaboration session to align the team on ethics: we combined industry-leading AI values with Zillow's core values and our design principles, landing on four commitments: fairness and integrity, inclusiveness, accountability, and transparency, with fair housing compliance woven through all of them. Those principles weren't a poster; they became acceptance criteria for the design decisions that followed. And I brought the content strategist in from day one, because in a product where the interface is mostly language, content strategy is half the design team.

AskZ AI principles artifact: Fairness & Integrity, Inclusiveness, Accountability, and Transparency, each paired with a Zillow core value and a working definition.
The principles, as workshopped. Each commitment paired with the Zillow core value it grew from. Fairness and integrity lead, with fair housing compliance woven through all of them. These became acceptance criteria for design decisions.
In an LLM product, the model's behavior is the material. I had to learn its grain the way a furniture maker learns wood.

Two directions, three weeks

Leadership gave us a deadline of just a few weeks for design direction, so I presented two: a traditional chatbot pattern, and an emerging AI-native UI. We chose the latter, stripped to what one Microsoft designer memorably called "deceptively simple": a clean, brand-aligned conversational interface where responses stay clear and concise, cutting cognitive load while keeping warmth. Simple to look at; very hard to keep simple.

Two early design directions for AskZ: a traditional chatbot entry with an inline Chat now call to action and popular-question chips, and an emerging AI UI with a floating chat button opening a full welcome screen with tips and a still-learning disclaimer.
The two directions, as presented to leadership. Left: the traditional chatbot, with an inline entry point and suggested-question chips. Right: the emerging AI pattern, with a floating entry and full-screen welcome with expectation-setting and legal slots. The bracketed [Intent] annotations are the content strategy collaboration in plain sight: design specified the job of each message, content owned the words. Elements of both directions survived into the shipped beta: the welcome tips from one, the suggested questions from the other.

The transparency principles showed up as concrete patterns: the product clearly disclosed that you were talking to an AI, acknowledged that answers could be wrong with avenues to verify, named its sources explicitly, and was upfront about data use. In 2023 these weren't yet table stakes; we assembled them from first principles and early industry practice.

Three AskZ beta screens: a welcome state that sets expectations, an answer with inline numbered citations and a Sources dropdown, and the feedback flow with report categories led by Offensive/discriminatory.
The shipped beta, in three states. Left: the welcome screen sets honest expectations, including that every question helps the model improve. Center: answers carry inline numbered citations with a Sources dropdown, plus per-response rating. Right: the reporting flow behind a thumbs-down, with “Offensive/discriminatory” deliberately first in the list. The trust loop, end to end.

Scope as a safety decision

The first real design decision wasn't a screen. It was the boundary. We deliberately launched finance-only. In a domain where a wrong answer can distort a six-figure decision, a narrow product that's reliable beats a broad one that's impressive, and every capability we declined to build was a failure mode we didn't have to design for.

The beta then taught us something important about boundaries: users don't read them. Many testers asked general home-buying questions anyway, which meant we either had to expand the knowledge or make the scope unmistakable. That finding directly shaped the scope-teaching patterns, like recommended next questions that quietly demonstrated what the product was good at, and informed the roadmap toward the broader home buying and selling scope that came next.

Fair housing: where AI safety got real

Real estate has a failure mode most AI products don't: the law. An LLM asked the wrong question about neighborhoods can walk straight into illegal steering: influencing where people look based on characteristics protected under the Fair Housing Act. And the stakes aren't abstract: Zillow's own research found a majority of Americans report experiencing housing discrimination in their lifetime, with far higher rates among LGBTQ+, Black, and Hispanic respondents. We spent a great deal of the project weaving fair housing guardrails into the experience, and the team developed a Fair Housing Classifier to detect noncompliant language, which Zillow later open-sourced.

My role sat around the classifier, not inside it. I was in the room as VPs and directors worked through why these guardrails had to exist, and I could contribute meaningfully because I'd been a real estate agent. Fair housing law wasn't a learning curve for me; it was a professional obligation I'd personally carried. Few designers in AI have operated under the regulations their product now has to respect.

From there, I contributed in two concrete ways. First, I helped train and test the model by deliberately asking it inappropriate questions, probing for the steering, bias, and boundary-pushing a real user (or a malicious one) would eventually try. I later learned the industry calls this red-teaming; at the time I just knew which questions an agent would get sued for answering. Second, when the classifier flagged a violation, the product still needed to respond: declining without shaming, explaining without a legal lecture. I designed those states and connected my content strategist to the legal and fair-housing parties needed to formulate the actual response rules, so the refusal language was correct, humane, and owned by the right people.

A classifier decides that the system won't answer. Design decides how it won't answer, and whether the person walks away trusting the product more or less for it.
In the pressZillow open-sourced the Fair Housing Classifier in 2024 so the wider industry could build on it. The release drew coverage in trade press, and the National Fair Housing Alliance's chief responsible AI officer publicly praised the open-source approach as a precedent for responsible innovation. → Zillow announcement · HousingWire coverage

Designing the failure states

An AI product is defined by its worst answer, not its best, so we treated the unhappy paths as core product surface:

  • Unknown answers. When the model didn't know, it said so, and when the question warranted a human, we linked directly to customer service rather than letting the model improvise.
  • Policy violations. Distinct, carefully written states for questions the product wouldn't answer, designed to preserve dignity and offer a path forward.
  • Time-outs and lost connections. Mundane failures, designed with the same care, because to the user, a hung response and a wrong response erode the same trust budget.
  • Grounding every answer. Responses signaled their sources with links to Zillow's own articles and listings, so people could verify rather than take the model's word for it.
AskZ user flow diagram showing the chatbot ingress, the knowledge-base decision point, a fully designed error branch with customer service escalation, a thinking state, answer population with feedback, and deep links to referenced resources.
The flow, with the failure branch drawn at full weight. The unknown-answer path gets equal structural treatment: a specific error message, suggested new questions, and a customer service escalation annotated directly on the node. Negative feedback routes to recovery, not a dead end. And yes, the strikethrough notes are real. Iteration leaves marks.

Closing the loop

Trust also needs an exhaust system. We built thumbs up/down on responses, with a reporting flow behind the thumbs-down so problems came back to us with enough detail to act on. And recommended next questions, related to what the person just asked, did double duty: they kept the conversation moving and quietly taught people what the product was good at, keeping usage inside the scope where we could be reliable.

Alongside our engineering partners and SLT executives, I participated directly in testing and training the model, from adversarial probing to evaluating everyday outputs and feeding failure patterns back into both the model and the interface. Design critique and model evaluation turned out to be the same muscle.

How we tested: dogfood, then the real world

Our primary launch goal was risk mitigation: get a high-quality experience in front of the broader public, not a flashy one. So we tested in rings. First, an internal MVP: we recruited across Zillow, had colleagues use the test environment, leave feedback on responses, and complete a survey, with our researcher synthesizing results. We learned a lot and shipped quick fixes before any external user saw the product.

Then a structured public beta: 10 days, 75 external users with the beta activated on their real Zillow accounts, measured on overall experience, relevancy, perceived latency, ease of use, value, and likelihood to use, followed by a six-person focus group to go deeper than the numbers.

The beta gave us two clear improvement paths: the scope confusion above, and the finding that responses were too text-heavy, pointing toward embedded objects like data visualizations and calculators as the way to make answers genuinely useful rather than merely correct. Testers loved what worked: 24/7 availability and fast answers to simple questions. That signal is ultimately what unlocked the investment that followed.

Outcome

The beta launched smoothly, which for a first LLM product is itself evidence the failure-state work paid off. The results led leadership to greenlight a full-service Zillow chatbot, since evolved into an AI copilot with its own dedicated design team of four, and a full-knowledge beta reaching thousands of users. AskZ became the direct foundation for Zillow's AI Search experience today. The fair housing guardrail work grew beyond our product entirely: out of the combined effort, the engineers on my team created the Fair Housing Classifier, which Zillow open-sourced so the rest of the industry could build on it.

What I'd do differentlyI committed to a three-week turnaround for design direction, and I paid for it: I sacrificed quality for speed and should have quoted a realistic timeline. It's a mistake I only needed to make once: I now scope AI design work with the honesty I expect from the products themselves.
Why this matters to meThis project changed how I think about the discipline. Designing for probabilistic systems means designing the refusals, the unknowns, and the boundaries as carefully as the happy path, and treating user trust as a budget you spend and earn rather than a brand attribute you declare.