Most OEMs don’t fail at AI because of technology. They fail because of organization.

In Part Two of our MAPconnected Service & Warranty Lifecycle Network Summit interview, Arnon Shafir, VP of Business Development at Upstream Security, pulls the conversation out of theory and straight into the hardest part of transformation: making AI real inside the enterprise.

This isn’t about “adding a tool.”

It’s about changing how quality, warranty, and aftersales teams measure success.

Arnon breaks down:

Why deploying AI demands new KPIs — accuracy, precision, recall — not just throughput and case closure

Why AI performance must be evaluated alongside human decision-making, not separately

How board-level mandates are shifting from experimentation to measurable financial impact

Why leading OEMs are targeting 5–15% reductions in warranty costs, based on joint research with McKinsey

How SDVs, edge computing, and in-vehicle HPCs — powered by companies like NVIDIA and Qualcomm — are redefining where quality decisions happen

Perhaps the most important signal he leaves us with:

Curiosity is the real leading indicator.

This is a rare inflection point for the industry — where AI-driven aftersales quality can emerge as an entirely new discipline. It won’t be unlocked by perfection, but by leaders willing to challenge assumptions and rethink quality from the boardroom down.

🎥 Watch Part Two of the interview and see why aftersales quality is no longer a downstream function — and why the biggest gains are still ahead.