P · chapter

Customer Service, Support & CX

What it is: the post-sale, non-revenue side of the customer relationship — resolving the contact (across voice, chat, messaging, self-service), running the helpdesk/contact-center, designing the human↔AI handoff, recovering from failure — plus the experience-management discipline (CXM/VoC) that measures delivered experience and closes the loop. It is also the home of AI service agents, the flagship shipped agentic use case in all of marketing.
The 2026 shift: the commercial model is flipping per-seat → per-resolution/outcome pricing, and autonomous resolution agents are GA, funded, and consolidating (Salesforce→Intercom Fin ~$3.6B, 2026-06-15, ledger). This is the most-shipped agentic workflow in the map.
The hard truth (P's defining tension): aggressive AI service erodes the loyalty service exists to build. The most credible independent research finds AI-for-service is the worst-performing measured AI use case, and 85% of service leaders are expanding human roles (Qualtrics + Gartner, 2026, ledger). The agent pole is the strongest in the map AND the human pole bites hardest here.
Don't: ship deflection as a solo metric, teach "engineer a failure to recover from" (the service-recovery paradox is folklore), trust any vendor resolution-rate as fact, or treat the AI→human handoff as something to minimize rather than design.

How this domain is organized internally — and the decision logic, runnable workflows, and agent angle — are in the full map.