I find the problem underneath the ask.

The question: how do Target's own food brands win a bigger share of a growing category? A month ago this was a hunch. Now it's mine to prove or disprove, end to end.

I tested whether the problem was real by querying the transaction source-of-truth tables directly, working with the teams closest to the data and layering in guest research. What surprised me: three core brands sounds narrow, but it splits across ~2,000 stores, in-store vs. delivery vs. order pickup, brand × item class, and how guests weigh value when budgets are tight.

Two root causes keep surfacing: reliability (is the item on the shelf, and is it still the item guests remember?) and visibility (do guests even see it when they're deciding?). Next: off-platform engagement through social, messaging and email, plus in-store pilots on how these items show up. Build starts in January.

The ask: “Can we use ChatGPT to evaluate an idea?” The real problem: every idea deserved a first review, but senior leaders' time is scarce, and most ideas were disappearing into documents and conversations.

I built the first evaluator prototype, refined it with Target-specific criteria, validated it with 100+ senior leaders and presented it to the CIPO and CGO. It shipped as an API inside Idea Workspace, where it is the first gate for every idea and a second opinion for human reviewers: 1,200+ ideas processed so far.

I own Idea Workspace's vision, roadmap and prioritization with engineering, and shaped the four-gate Idea Pipeline around it: AI review, leader decision, refinement, then build, incubate or stop with the learning written down. 1,500+ ideas across 50+ campaigns. The point is moving ideas forward, not collecting them.

The ask: run a week-long live onboarding program. The real problem: there was no onboarding system at all. New hires, hiring managers and buddies across India and the US were each improvising.

I treated it as a product with three users. I mapped a journey for each, aligned HR, hiring managers and tech teams in both countries, and replaced the live week with a semi-autonomous five-week journey: migrated content, guided walkthroughs and AI chat for questions. It now serves 500+ tech teams, with an NPS above 40 and over 90% of users rating it positively. The answer was a system, not another app.

The problem: people knew about the GenAI tools. They weren't using them, and more awareness wasn't going to change that.

So I helped turn adoption into hands-on building: Target's largest GenAI hackathon, with 1,400+ team members, ~200 apps in one week and the highest weekly internal AI-platform adoption on record at the time. The number I cared about came after: ~20 teams kept building.

Around it, I built deliberately small tools for narrow problems: AI Fluency Pathways (1,200+ users) and personalized Slack outreach that lifted conversion by 300–400%.

Early at Target, I joined a product fellowship with supply-chain partners. A forecasting capability already existed, but planners waited days for another team to run it. The product was mostly built when I joined; I helped finish the last pieces and took it to its first users.

Rollout taught me what a launch plan doesn't: the questions real users ask on day one, how internal adoption actually happens, and how to turn that into the next iteration's requirements. Turnaround dropped by up to 90%.