Part 3 — Quality & Measurement

Part 3 overview from The AI Contact Center Handbook by Sho Shimoda. Available on Amazon.

Why Part 3 exists

Parts 1 and 2 introduced the landscape and the front line — the architecture of AI-driven contact centers and the agents (human, AI, and blended) who staff them. Part 3 turns to the question that follows immediately after any operating team gets those two things running: how do you measure quality when the agent doing the talking is not the human you're evaluating?

The old quality model was straightforward. A supervisor listened to a sample of calls, scored the human agent against a rubric, coached them, and moved on. The scorecard was the artifact. The human was the subject. Everything the human did, they did themselves.

None of those assumptions hold when an AI is drafting the reply, a bot is handling the first two minutes of the call, an orchestrator is deciding which channel the customer lands on, and a translation engine is doing the actual speaking. The subject of quality shrinks — the human is one participant among many. The scorecard has to grow — it needs to score the AI's contribution, the orchestrator's routing decision, the personalization engine's targeting choice, and the composite experience the customer actually had.

Quality in the AI era is not a supervisor scoring an agent. It is a system scoring itself while a human decides what "good" looks like.

The two chapters in Part 3

Chapter 6 WEM Chapter 7 Personalization Score the work. Recognize the customer. Two halves of the same problem: does the system know what happened?
Figure — the arc of Part 3.
Chapter 6
Intelligent Workforce Engagement Management (WEM)
Forecasting, scheduling, quality management, coaching, and gamification — reinvented for a workforce where half the agents are AI. From Erlang C to auto-QA on 100% of interactions.
Read Chapter 6 →
Chapter 7
Hyper-Personalization and Omnichannel Journeys
Why omnichannel broke, what optichannel replaces it with, live journey analytics, real-time speech-to-speech translation, and the trust line that separates delight from creepiness.
Read Chapter 7 →

What Part 3 buys you

If you are… Part 3 gives you…
A contact center operations leader A scorecard model that survives the shift to blended human-AI staffing, and a channel-cost lens that turns routing into economics rather than habit.
A CX or quality director A path from sample-based QA to 100%-coverage auto-QA, plus a way to score the customer's whole journey rather than one interaction at a time.
A marketing or CRM leader The vocabulary for zero-, first-, second-, and third-party data, and the rule that separates personalization the customer welcomes from personalization they resent.
An architect or product lead A picture of the CDP + orchestration + journey-analytics stack that has to sit underneath any of it working in real time.
A frontline manager A coaching model that uses AI to spot the moments that matter across every interaction, not the seven calls a week you could sample by hand.

What comes next

Part 3 assumes the doors are open — that the technology, the data, and the people are lined up to deliver a modern experience. Part 4 asks the opposite question: which doors have to stay closed? Regulation, security, privacy, bias, explainability, and the compliance obligations that reach into every AI decision a contact center makes. It is the part of the book that turns the enabling capabilities of Parts 1 through 3 into safely deployable systems.

Continue to Part 4 — Governance & Safety →

Read the whole book
The AI Contact Center Handbook
Twelve chapters, five parts, plus six practical appendices. Paperback and Kindle.
View on Amazon →
Quality in practice
AB Support for Microsoft Teams
A support desk built inside Teams — auto-QA, journey visibility, and coaching signals where the work already happens.
Learn more →

Published on: 2026-08-09 Last updated on: 2026-08-09

Questions & answers

Have a question about this topic? Ask below — no sign-up needed. The team reviews and answers questions here.

No questions yet — be the first to ask.

Ask a question

We’ll send a one-time email to confirm your address. Questions appear after a quick review.