Design a privacy-safe personalization system for Support at global scale

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 Support serves consumers, Dashers, and merchants across a complex local commerce marketplace. In this question, focus on grocery shoppers who may contact Support for issues such as missing items, substitutions, delayed deliveries, refunds, order accuracy, payment problems, or communication with shoppers/Dashers. The challenge is to design a personalization system that helps Support provide faster, more relevant, and more trustworthy experiences without compromising user privacy.

You are designing this as a Technical PM at global scale, where the system may need to work across markets, languages, regulatory environments, product surfaces, and support channels such as in-app help, chat, phone, email, and automated flows. Personalization could influence issue triage, suggested help content, agent context, proactive updates, escalation paths, or resolution options, but the design must clearly define what data is used, how it is protected, and how user trust is maintained.

The problem is not just to make Support “smarter.” You should frame the end-to-end support workflow, identify where personalization adds value, and specify the technical and product requirements needed to make the system reliable, safe, explainable, and operationally scalable. Consider the trade-offs between speed, accuracy, privacy, compliance, support cost, customer satisfaction, merchant impact, Dasher impact, and abuse prevention.

The experience should consider:

- The primary users and stakeholders: grocery shoppers, support agents, automation systems, Dashers, merchants, and internal operations teams.

- The support journey from issue detection or user contact through triage, context gathering, resolution, follow-up, and escalation.

- What types of personalization are appropriate, what data they require, and what data should be excluded or minimized.

- APIs, data pipelines, identity/session handling, consent, retention, access control, auditability, and cross-system dependencies.

- Privacy, security, compliance, fairness, explainability, and regional differences in data handling requirements.

- Reliability expectations, fallback behavior, latency, uptime, model/service degradation, and human-in-the-loop workflows.

- Rollout strategy, experimentation, monitoring, incident response, abuse prevention, and observability.

- Product trade-offs across user experience, operational efficiency, marketplace balance, unit economics, and trust.

Your goal is to describe a technically credible product design for a privacy-safe personalization system that improves DoorDash Support outcomes for grocery shoppers at global scale, while making clear what must be built, how it integrates with existing systems, how it protects users, and how the team would evaluate whether it is working responsibly.