How Zillow’s AI Strategy is Redefining Real Estate Transactions

In the fast-evolving landscape of real estate technology, Zillow stands out not just for its market presence but for its innovative approach to artificial intelligence (AI). At the recent VB Transform 2026 conference, the company’s Senior Vice President of Engineering, Toby Roberts, highlighted a critical insight: the true return on investment (ROI) from AI initiatives is contingent upon establishing a robust measurement framework prior to deployment. This principle has guided Zillow’s journey, where the focus is less on the data itself and more on the contextual understanding of customer interactions throughout their real estate journey.

The traditional real estate process is often convoluted. Buyers and sellers may engage with various touchpoints, from websites and mobile apps to loan officers and real estate agents, sometimes spanning months or even years. In this fragmented journey, maintaining context is key. A simple chatbot cannot carry the thread of conversation across multiple interactions, which is where Zillow’s advanced AI architecture comes into play.

The Importance of Context in Real Estate Transactions

Zillow’s AI initiatives impact approximately 80% of U.S. real estate transactions annually, highlighting the scale at which it operates. According to Roberts, rather than merely processing data, the company recognized the necessity for a persistent context layer that would seamlessly support users across different stages of their journey. This shift in focus has proven to be a more complex challenge than simply aggregating and analyzing data.

Understanding Customer Journeys

Roberts explained that Zillow's initial AI efforts were rooted in establishing a strong data foundation. This foundation relied on a data mesh approach, which emphasizes decentralized data ownership and clear data lineage, coupled with stringent governance protocols. However, the real challenge emerged when the team sought to create a system capable of remembering a user's progress and preferences, no matter where they engaged with Zillow’s services next.

• Persistent Context Layer: Allows for seamless transitions across different stages of the real estate journey.
• Data Mesh Approach: Decentralizes data ownership to enable better integration and governance.
• Customer-Centric Design: Focuses on understanding and anticipating user needs throughout their journey.

Building a Tailored AI Architecture

Instead of relying on a generic AI model, Zillow opted to build its own specialized architecture. This decision stemmed from the recognition that a singular model wouldn’t suffice for the varied tasks required in real estate transactions. Drawing from over two decades of experience in machine learning, Zillow has implemented a system that utilizes smaller, task-specific models rather than a one-size-fits-all solution.

Integration with Glean

A significant aspect of this architecture is its integration with Glean, a platform designed to streamline AI interactions. Roberts noted that Zillow has deployed thousands of Glean agents internally, which handle repetitive tasks and improve operational efficiency. Glean's model centralizes integration efforts across various departments—finance, legal, and marketing—preventing redundant work and hidden costs associated with duplicated integrations.

Strategies for Cost Efficiency in AI Deployment

One of the standout features of Zillow’s AI strategy is its focus on cost efficiency. Arvind Jain, co-founder and CEO of Glean, articulated two key mechanisms aiding in this endeavor:

• Model Routing: This method directs tasks to smaller, more cost-effective models instead of defaulting to more expensive ones.
• Precomputed Context: By assembling context ahead of time, Zillow avoids the inefficiencies of agents having to create context from scratch for each interaction.

Reducing Token Consumption

These strategies contribute significantly to reducing operational costs. For instance, Jain pointed out that using Glean for request processing can cut token consumption by as much as 50%, which is crucial in an environment where API calls can become costly. This focus on efficiency isn’t just about saving money; it’s about enabling more robust and responsive customer interactions without overspending.

Lessons for Enterprises Embracing AI

The insights shared by Roberts and Jain provide valuable lessons for enterprises looking to implement AI systems in their operations. Here are some critical takeaways:

• Establish Measurement Baselines: Before rolling out AI solutions, companies should create a baseline for performance metrics. Zillow’s attribution of a 40% increase in shipped code to AI adoption was built on a solid DORA metrics foundation established long before their AI initiatives.
• Centralize Context Management: Avoid redundant efforts by creating a unified context management system that can be leveraged across departments.
• Implement Rigorous Data Governance: Just having a permissions-aware architecture isn’t enough; companies must layer compliance checks on top of sensitive data.
• View Context as a Cost Lever: By treating context management as a critical component of cost efficiency, organizations can maximize the impact of their AI investments.

Key Takeaways

• Context in AI is crucial for enhancing customer experiences and operational efficiency.
• Zillow’s focus on a persistent context layer has set it apart in the real estate tech landscape.
• Cost efficiency strategies, such as model routing and precomputed context, can significantly lower operational expenses.
• Establishing measurement baselines is essential for justifying AI investments.
• Centralizing context management prevents duplicated efforts and hidden costs.

Frequently Asked Questions

What is Zillow's approach to AI in real estate?

Zillow has developed a unique AI architecture that prioritizes the creation of a persistent context layer, allowing for seamless transitions between different stages of the real estate transaction process. This approach helps maintain continuity for users and improves overall customer experience.

How does Glean fit into Zillow's AI strategy?

Glean is an integration platform that centralizes AI interactions and processes within Zillow. By using Glean, Zillow can deploy thousands of specialized agents to handle repetitive tasks, thus enhancing efficiency and reducing operational costs through methods like model routing and precomputed context.

Why is measuring AI ROI important?

Measuring AI ROI is crucial because it provides enterprises with a clear understanding of the impact of their AI initiatives. Zillow emphasizes establishing measurement baselines before deploying AI solutions, allowing them to credibly attribute improvements in performance to specific AI investments.

What are some common pitfalls for enterprises adopting AI?

Common pitfalls include failing to create a measurement baseline before AI deployment, allowing duplicated integrations across teams, and not implementing rigorous data governance for sensitive information. These issues can lead to inefficiencies and hidden costs that undermine the potential benefits of AI technology.