Part 6 — Scaling, Measuring, and Governing Agentic AI

Overview

Most organisations can run a successful pilot. Far fewer can turn that pilot into something that runs reliably at scale, justifies its cost, survives operational reality, and becomes embedded in the organisation rather than withering when the initial sponsor moves on.

Part 6 addresses the gap between a proof of concept that impresses in a demo and a production capability that generates durable value. This gap is not primarily technical. The agents that stall between pilot and production almost always do so for organisational, commercial, measurement, governance, or human reasons — not simply because the model underperformed. Understanding those reasons, and designing deliberately to address them, is what this part is about.

The four chapters form a deliberate arc: scale, measure, govern, and sustain. Chapter 20 examines the structural obstacles that separate successful pilots from scaled production systems: funding transitions, sponsor risk, operational handoff, and sequencing decisions. Chapter 21 addresses the measurement problem: how to build a credible business case, track value in production, manage model updates, and keep deployed agents improving rather than silently degrading. Chapter 22 turns to human oversight and governance design: how to make human review meaningful, preserve skill development, calibrate trust, and build audit trails that support accountability. Chapter 23 closes the part by addressing the capability question: what roles, partners, platforms, and learning structures are needed to sustain agentic AI beyond the first wave of deployments.

Together, these chapters make the practical argument that production agentic AI is not a one-time implementation. It is an operating discipline. The organisations that succeed will not be those that simply launch more agents, but those that build the measurement, governance, human capability, and team structures required to keep those agents reliable, valuable, and accountable over time.

---

Chapters in This Part

Chapter	Title	Theme
20	Crossing the Valley: From Successful Pilot to Production Scale	Scaling dynamics
21	Measuring and Improving Agent Performance	Performance and value measurement
22	Human Oversight and Governance Design	Oversight and accountability
23	Building the Team: People, Partners, and Platforms	Capability building

---

These four chapters form a continuous arc and are best read in sequence. The problems they address tend to arise in this order in practice: scaling first, measurement next, governance as autonomy increases, and sustained capability once the programme becomes part of the organisation.