A leader who uses AI to prepare a decision has already made a smaller decision first: how much weight the output will carry. Chapter 7 described decision quality without AI. This chapter asks what changes when an assistant supplies the options, drafts the paper in which they are framed, or is permitted to take a routine decision on its own. It concerns L2 Decide, the function of taking decisions of quality and of deciding who decides, and it lands on L5 Answer, because every degree of delegation described here ends in the question of who can be named, who can explain and who can reach the decision to stop it.
The chapter separates three roles that AI can hold in a leadership decision (adviser, drafter and decider) and says what record each requires; reports what studies measured about reliance on machine advice, including results that do not fit a simple story; sets out how a leader calibrates trust and how the delegation ladder of Chapter 4 works for narrow routine decisions; and stages the debate on whether leaders should let AI decide, with a verdict by decision class.
Four conclusions of the earlier volumes frame the chapter. Agents draft and humans approve: in Architecture as Code (Chapter 20) AI agents are structured participants under human direction, with a named human approval for exceptions and risk acceptances (220, Nordqvist, 2025). Oversight is designed, not declared: AI Governance (Chapters 10 and 21) required an interruption path, an override log and the authority to use them, and read approvals given in seconds as rubber-stamping (221, Nordqvist, 2026). Autonomy has levels, and the level is decided and recorded: AI Red Team (Chapter 6) set Level 0 Advisory to Level 4 Autonomous, with Level 3 defined by a codified playbook enforced by a policy decision point outside the model (223, Nordqvist, 2026); AI Operations (Chapter 4) made the level the lowest of four answers (reversibility, blast radius, evidence and window) (229, Nordqvist, 2026). And decisions about people differ in kind: AI Boss (Chapters 4 and 33) concluded that AI may direct work but may not, on its own, decide about a person (226, Nordqvist, 2026). AI Threat (Chapters 6 and 14) adds the warning behind this chapter: reliance on AI reduces effort, recall and, among those who trust the tool, checking (224, Nordqvist, 2026).
Primary topics explored
- What the series has already settled
- Adviser, drafter, decider
- What studies measured about reliance on machine advice
- Calibrated trust
- The ladder in use
- The debate: should leaders let AI decide?
About AI Leader
AI Leader is Book 11 of The Transformed Architecture. It takes leadership as its subject and asks what happens when AI is used broadly: the five leadership functions (orient, decide, organise, enable, answer), the delegation ladder and the Level 4 prohibitions, the answerability principle (name, explain, reach), the span and the middle layer, human-agent work, the board and the management body, change and adoption as claims that need evidence, regulation (the AI Act, DORA, NIS 2, the GDPR), ethics and sovereignty, and the leader as a person, for European enterprises and public authorities.
It builds on the conclusions of Architecture as Code, AI Governance, AI Platform and Security, AI Red Team, AI Threat, The Transformed Security Architect, AI Boss, AI Blue, AI Purple and AI Operations.
chairs and members of boards, chief executives and executive teams, the changing C-suite, middle managers and first-line leaders, heads of public authorities and the politicians who oversee them, HR specialists, works councils and unions, architects who advise leaders, and supervisors and auditors.
Gunnar Nordqvist
Conceived, directed and published the books of The Transformed Architecture.
Continue with the full chapter
Chapter 8: Decide with AI: Advice, Automation Bias and Delegated Decisions in AI Leader gives the detailed explanations, practical examples and implementation patterns behind this summary.
This whitepaper belongs to “The Craft of Leading” and is one of 37 drawn from the manuscript.