Building a platform that surfaces global Ads customer feedback at scale.
This work is under NDA. Visuals and specific metrics are redacted. The narrative focuses on design thinking, strategic approach, and impact at a structural level.
A 0 → 1 initiative to build a platform from the ground up that surfaces global Ads customer feedback at scale. The tool enables product teams to identify high-friction pain points and market opportunities in real-time — think Sprinklr, but purpose-built for Google Ads.
Each Ads product team operated in silos, running ad-hoc reactive UXR sessions that were expensive and created significant bottlenecks in rapid product iterations and GTM adjustments. Teams maintained custom dashboards to track usage behavior. Each function — market leads, verticals, sales programs — tracked metrics from their own vantage point. Cross-pollination of insights happened organically in select places, but never at scale. There was no ability to form narratives around macro-trends. When ChatGPT entered the market and began affecting advertiser behavior, spend, and creative asset creation, there was no unified lens to understand the shift and adapt.
Partnered out-of-band with L7 leaders to co-develop the platform vision. Showcased the concept to Google’s CBO to secure project funding.
Designed the frameworks for how Gemini-driven insights are categorized and visualized for product stakeholders. The project surfaced unique challenges: dynamically generated insights did well at identifying themes, but their fluid nature meant there was no stable measurement baseline for tracking trends over time. The solution started with business-led theme buckets — listening for specific signals, then layering human and auto-raters to assess quality and relevance.
This also meant solving for transparency: where does the data come from? If a user sees 50K positive mentions, what’s the total universe? Is that signal strong or noise? What about breakdowns by market, vertical, and sales program — which voices are loudest? And competitor mentions — are there positioning, product, or sales gaps we could close?
Led the creation of high-fidelity prototypes that served as the primary vehicle for demonstrating value and securing leadership buy-in.
Balancing dynamic AI-generated themes vs. curated theme buckets required careful design iteration to find the right mix of flexibility and consistency.
Balancing the needs for actionable insights and leadership reporting required careful calibration of truthfulness and usefulness over visual flair.
Initial approach of asking each product team for their data slices quickly risked enterprise bloat. Shifted to leveraging AI patterns and mental models — starting with a few stable filters, then enabling organic discovery through dynamically generative views and chat.
An ambiguous space where each persona required different treatments. Required extensive alignment conversations and change management across functions.
Perpetually in a state of beginner's mind.