The field library
Methods leaders can use, challenge and revise.
Short guidance for making consequential operating choices—not trend reports, model rankings or generic transformation theatre.
Authority before autonomy
Why agent autonomy is an operating permission—not a model capability—and what leaders must define before granting it.
Operating model · COO, business and transformation leadersRedesign work, not tasks
A practical way to move from scattered AI assistance to end-to-end agentic operations without automating organizational debt.
Control brief · CIO, CTO, CDAO, risk and auditGovernance at machine speed
Replace periodic approval with controls that travel with agent identity, context and action.
Board brief · Board and CEOA board agenda for agentic operations
Five questions that move board oversight from AI activity reports to strategy, authority, operating evidence and resilience.
Agent
A software system that pursues a goal through a sequence of decisions or actions with some operational autonomy.
Agency
Permission and capability to affect an environment—not simply generate information.
Operating envelope
The authorized scope, context, tools, limits, conditions and escalation rules governing action.
Override authority
The named human authority able to pause, reject, redirect or withdraw delegated action.
Control floor
The least mature material system that limits how much agency can safely be granted.
Selected source notes
A foundation to examine—not an appeal to authority.
- 01OECD · The Agentic AI Landscape and Its Conceptual Foundations (2026)
Autonomy, goal-directed action and interaction with environments; accountability, explainability and continuous-monitoring challenges.
- 02NIST · AI Agent Standards Initiative (2026)
Interoperability, secure human-agent interaction and the emerging standards infrastructure for trustworthy agent systems.
- 03McKinsey · The agentic organization: A new operating model for AI
The connection between business model, operating model, governance, workforce, technology and data; real-time embedded governance.
- 04BCG · Reinventing the Operating System of Work with AI (2026)
Moving the unit of transformation from isolated tasks to end-to-end processes under defined controls and accountability.
- 05Bain · Agentic AI governance, risk and controls
Enforceable action and resource limits, containment, approval for consequential actions and tested rollback.
- 06NIST · AI Risk Management Framework
A durable baseline for governing, mapping, measuring and managing AI risk.