Head of Product · Retail-Tech / Fintech · Current product
Purchase Intelligence
An AI-powered insights product built on top of large-scale purchase and receipt data, helping merchants understand customer behavior and unlock new revenue streams.
Context
anybill sits on one of the richest first-party purchase datasets in retail — receipts at the line-item level across thousands of merchants. Until recently, that data powered compliance and convenience. The opportunity was bigger: turn it into a product merchants actively use to grow.
Purchase Intelligence started as a hypothesis about expansion revenue. Could the existing customer base unlock a second product — one that monetizes the data layer rather than the receipt itself? The work began with merchant interviews, sharp scoping, and a prototype built outside the existing roadmap.
The product had to land in a market shaped by retail-tech buying cycles, regulated data handling, and very different levels of analytics maturity across customers. Simplicity, defensibility, and a clear "so what" were non-negotiable.
What I did
- Owned the product from problem framing through 0→1 build, working hand-in-hand with engineering, data, and commercial teams.
- Designed an AI-powered insights layer on top of the existing purchase and receipt data, prioritizing use cases with clear merchant ROI.
- Aligned product, sales, and partnerships behind a focused commercial wedge inside the existing customer base.
- Set up the discovery and delivery rhythm — outcome-focused roadmap, measurable bets, and weekly merchant feedback loops.
- Connected the product to broader anybill platform decisions on pricing, packaging, and partner integrations.
Outcome
A new revenue stream inside an existing customer base — measurable expansion revenue, sharper positioning in retail-tech, and a product line that earned its place next to the core platform.
Expansion revenue growth
What I learned
In B2B SaaS, "AI features" are not a strategy. The leverage is in choosing the one merchant problem where AI changes the unit economics, and then defending that wedge with data quality and tight feedback loops.