Design a privacy-safe personalization system for Support

Problem Statement Description

Product context: DoorDash is a local commerce and delivery platform; its products include restaurant delivery, DashPass, grocery and retail delivery, merchant tools, and dasher tools.

DoorDash wants to improve the Support experience for grocery shoppers by making support interactions more relevant, faster, and context-aware while protecting customer privacy. Grocery orders have distinct support needs compared with restaurant delivery: substitutions, missing items, temperature-sensitive products, delivery windows, store availability, promotions, and larger basket values can all create more complex help scenarios.

In this technical PM interview, you are asked to design a privacy-safe personalization system for DoorDash Support. The system should help customers, support agents, and automated support flows use appropriate context without exposing unnecessary personal data or creating unfair, opaque, or unsafe outcomes. You should think through how personalization works across in-app help, chat, agent tooling, escalation paths, and post-resolution follow-up.

Your scope should include the product requirements, data and system interfaces, consent and privacy controls, reliability expectations, observability, and rollout considerations. The focus is not only what the support experience should feel like, but also how the underlying technical system should safely decide what context is available, when it is used, and how its impact is measured.

The experience should consider:

- Grocery shopper workflows before, during, and after a support contact, including order issues, refunds, substitutions, delivery delays, and repeat problems

- What user, order, merchant, Dasher, and support-history data may be useful, and what should be minimized, masked, aggregated, or excluded

- API and data requirements for integrating personalization into customer-facing help flows and support-agent tools

- Privacy, consent, retention, access control, auditability, and security expectations for sensitive customer and order data

- Reliability and fallback behavior when personalization signals are missing, stale, incorrect, or unavailable

- Guardrails to avoid biased treatment, over-personalization, inappropriate automation, or inconsistent resolutions

- Rollout, experimentation, monitoring, and incident response plans for a support-critical system

- Trade-offs between faster resolution, customer trust, operational cost, marketplace fairness, and regulatory risk

The goal is to define a technically credible product system that improves grocery shopper support outcomes while maintaining DoorDash’s trust, privacy, and marketplace reliability standards.