Work

Five case studies, in the order a traveler meets them, then shorter notes on more of the work. Results are team results unless I say otherwise, and every number says where it comes from.

More work

  • Hertz, 2026

    Helping renters extend before they go overdue

    I led Hertz’s program to help renters extend before they go overdue, starting with nationwide pre-return reminders that link straight to self-service extension. When we launched overdue reminder texts, a targeting error the next day reached renters who weren’t overdue. I owned the root-cause analysis within 24 hours, moved eligibility to the system of record with tested kill switches, and relaunched cleanly. Renters who got the texts extended about 5 points more often. Hertz’s AI agent now also contacts renters about their return plans, work I owned with Decagon.

    Directional. There was no holdout group.

  • Hertz, 2026

    One flight-status service for every channel

    I owned one shared flight-status service that Hertz’s AI chat, AI voice and customer messages build on. During 2026 it went from APIs, to live data streams, to the AI channels, across Hertz, Dollar and Thrifty. Before that, fixing our flight-delay text taught us that reach is not deflection: the text reached nearly three times as many renters, which said nothing yet about whether fewer of them called.

    No impact claim. Reach is not the same as fewer calls.

  • Hertz, 2024 to now

    The human contact-rate program

    I own the human contact-rate goal for Hertz’s US customer care across Hertz, Dollar and Thrifty. I broke one lagging ratio into root causes and reduction tiers, and told leadership plainly what the plan could and could not deliver. The human contact rate fell about 11% from 2024 to 2025.

    Program-level, with AI as one lever among several.

  • Hertz, 2026

    AI chat and voice for Dollar and Thrifty

    I extended Hertz’s AI chat and voice to the Dollar and Thrifty brands by reusing the proven Hertz setup. Voice ramped to every call within days, with every step gated on operations sign-off and a one-click off switch. I also wrote the US blueprint for international teams bringing self-service rental extension to their markets.

  • Hertz, 2026

    Repeat callers, starting with a clean metric

    Before trying to cut repeat calls, I cleaned the metric: shared partner phone lines were inflating repeat counts. The analysis led to a fix that gives Hertz’s AI agent memory of a caller’s prior calls. It shipped in the second half of 2026.

    Impact not measured yet.

  • Hertz, 2026

    Finding the missing cancellation emails

    Some customers never got a cancellation email. A strong correlation pointed at the wrong cause. I checked what the data field really meant with the team that owns it, confirmed the cause from a second source, and found it: cancellations made by phone or through an agent never triggered the email. I also built a tracker for every transactional email and text type, with every unvalidated view labeled as a draft.

  • Hertz, 2024 to 2025

    Fixing self-service dead ends

    I took on the unglamorous reliability work behind changing a rental yourself. In my first months, rental-extension and booking-lookup errors fell sharply and more customers extended digitally. I also fixed dead ends that pushed customers to call, such as showing a clear “cancelled” status instead of a misleading “not found” error.

  • Hertz, 2026

    Making customers reachable

    I shipped phone verification across the app with a tiered design: required for premium bookings, optional elsewhere, and switchable off in one step during an incident. I split the reachability problem into two fixes: capturing contact details on bookings made through partners, and verifying phone numbers everywhere.

  • Hertz, 2026

    Acting product lead, Skip the Counter

    I served as acting product lead for Hertz’s Skip the Counter airport check-in product, keeping executive reviews and expansion work moving.

  • Walmart, 2022 to 2024

    Last-mile delivery on Spark

    As a Principal Product Manager at Walmart, I worked on last-mile delivery by gig drivers on the Spark platform: Spark Shop & Deliver, order batching and a drone delivery pilot. Across my Walmart work, the programs delivered about $750M in combined savings and revenue.

    As I reported it at the time.

  • Walmart, 2022 to 2024

    Drone and bot delivery

    I worked on the strategy and roadmap for faster q-commerce delivery by drone and bot. That meant user research, a pilot with technology partners and the success metrics.

  • Walmart, 2022 to 2024

    Add Item Post Order Placement

    I worked on Add Item Post Order Placement for Walmart Ecommerce, which lets customers add forgotten items after they order. It contributed to a $400M+ annual GMV increase.

    As I reported it at the time.

  • Allstate, 2018 to 2020

    Liability Analysis Tool

    I delivered Allstate’s Liability Analysis Tool for more than 2,000 claims adjusters, cutting liability settlement time from about 30 days to hours.

    As I reported it at the time.

  • Allstate, 2021

    Quick Foto Claim, modernized

    I modernized photo intake for Allstate’s Quick Foto Claim, which handled more than 600,000 photo packages and 6 million images by November 2021. Allstate first launched the feature in 2013.

    As I reported it at the time.

  • Allstate, 2020 to 2021

    Sales agent platform, from idea to MVP in six weeks

    I took Allstate’s sales-agent platform from idea to a minimum viable product in six weeks. Within nine months, more than 400,000 customers had expressed interest, and an onboarding A/B test lifted conversion about 10%.

    As I reported it at the time.

  • Allstate, 2020 to 2021

    CollectiveNow

    I worked on CollectiveNow, a crowd-sourced auto estimation platform where people earn income from their expertise. It cut the cost of an estimate by about two-thirds.

    As I reported it at the time.