Zillow Explains the Risks of Building AI ROI Without Measurement
Real-time context layer determines AI success.
1Zillow's AI journey
Zillow has come a long way in understanding AI's role in real estate transactions. Tracking customer context emerged as a more challenging problem than raw data. This highlights AI's need to focus on context, not just data.
2Importance of the context layer
Zillow developed a context layer that tracks customers' positions in the real estate process. This layer ensures context is maintained as customers transition between platforms. It's a critical step in unlocking AI's true business value.
3Glean's role
Zillow automated repetitive tasks by working with Glean. Glean's centralized integration approach streamlined connections for finance, legal, and marketing teams. This increased efficiency while reducing costs.
4Model routing and cost
Glean reduced token consumption by routing tasks to smaller, cheaper models. Precomputed context sped up the model's context-building process. This strategy is key to controlling AI costs.
5Importance of metrics
Zillow used DORA metrics to measure AI's impact, proving it increased code shipping speed by 40%. Metrics are a fundamental tool for evaluating AI success.
6The future of AI
Zillow and Glean's approach offers valuable lessons on implementing AI in enterprise systems. Context management and cost control emerge as critical factors in AI's success.
Original article: VentureBeat — Zillow, AI ROI'yi ölçmeden inşa etmenin risklerini açıkladı