loader image
Skip to content

Income & Data Decisioning Platforms

Case Study · 2023 to present

Product leadership for the income data behind credit decisions on 16M+ consumer accounts.

Every new card application and credit line increase depends on accurate income, housing, employment, and asset data. These platforms collect, store, and verify that data, and the verification logic helps set final credit limits.

16M+ consumer accounts supported by the platforms20% fewer open decisioning exceptions and a 30% lower true exception rate
Organization
Capital One, U.S. Card
Role
Senior Data/AI/Risk Product Manager
Dates
August 2023 to present
Skills
Platform strategy, data strategy, PRDs and roadmaps, AI governance, human-in-the-loop design, regulatory alignment (FCRA, ECOA, fair lending), executive communication.
The Challenge

Income data in a regulated credit business has to be accurate, explainable, and audit-ready, and every exception means manual review for risk and operations teams. New tools like generative AI promise speed, but only if accountability stays intact.

My Role

Product owner for three platform pillars: income storage, income collection, and verification of income. Serve as the go-to source on income data collection and data strategy for the U.S. Card division, and bring legal, risk, compliance, business, and engineering partners to one roadmap.

What I Did
  1. Step 1

    Led root-cause analysis and production logic fixes on decisioning exceptions.

  2. Step 2

    Directed a document-intelligence program that brings generative AI into card decisioning governance, with human reviewers built into every AI-assisted step.

  3. Step 3

    Managed recurring large-scale income data ingestion from external bureau partners, coordinating data engineering and compliance on quality and timeliness.

  4. Step 4

    Designed an intake and operating model with five delivery lanes and a zero-to-one intent flow with a named owner and stage SLAs for every job family.

  5. Step 5

    Leading the team's adoption of spec-driven development.

  6. Step 6

    Launched a monthly leadership newsletter and engineering deep dives on each initiative.

Results
16M+consumer accounts supported by the platforms
20%fewer open decisioning exceptions and a 30% lower true exception rate
  • Five-lane operating model now rolling out across teams.
  • Monthly newsletter reaching 300+ stakeholders.
  • Community & Events Lead for a 200+ member early-career product council; grew engagement 25%.
Ready to build something like this?

Let’s talk about your team.

Whether you’re building a data product, setting up AI governance, or growing a product team, I’d love to hear what you’re working on.