Better Customer Context, Better AI: Rethinking CX Automation: A Conversation with Ipsy's TJ Stein

Watch the Live Recording
Introduction
TJ Stein, Head of Customer Care at Ipsy, joins host Lisa Painter to talk about why customer context is the key to AI performance, and why the model itself matters far less than most CX leaders think. TJ walks through his three-tier AI maturity model, from glorified knowledge base to AI that takes action on a customer's behalf, and explains why repeat contacts about the same resolved issue are the context most teams miss. They also dig into the unglamorous groundwork that makes AI work (clear SOPs, governance, and dedicated owners), the misconceptions leaders still hold about plug-and-play AI, and the metrics TJ trusts to tell real value from simple automation.
We're still very much in early days, so I'm skeptical of anyone that says they've figured it all out. Everybody's still learning right now.
TJ Stein
Head of Customer Care, Ipsy
Key Takeaways
Build context in three tiers
AI maturity moves from basic knowledge base retrieval, to personalized answers like tracking and store credit, to AI that acts on a customer's behalf by processing returns and refunds. Each step needs deeper integrations with internal systems.
Treat every contact as part of a story
Most support systems treat each conversation as a standalone event, so the third contact about the same issue gets the same scripted apology. For the customer, that's strike three.
Expand context before chasing a better model
Ipsy's biggest gains came from sending more context into the platform and giving the AI more API endpoints to take action, not from model advances. They found the gaps by analyzing handoffs and frustration points.
Replace "it depends" with clear SOPs
AI struggles when a service model runs on judgment and discretion. Structured SOPs, escalation paths, guardrails, and clear ownership set the foundation.
Measure value beyond deflection
TJ looks at fully loaded cost per resolution, sentiment (CSAT, promoter, or effort score), and true resolution rate, all tied to business outcomes like retention and lifetime value.
Meet the Guest
TJ Stein is Head of Customer Care at Ipsy, the beauty subscription brand where every customer interaction builds on the last. He leads Ipsy's push to give AI deeper customer context, from order history to the ability to take action on a member's behalf, and he's candid that the work is harder than most vendors admit. His point of view: stay ruthlessly focused on great customer outcomes, no matter how noisy the AI landscape gets.



