AI-P08 — AI for Executive Decision-Making

Course Code: AI-P08  |  School: School of Artificial Intelligence  |  Cluster: Level 3 — Applied & Sectoral Practice

Level: Intermediate  |  Duration: 5 weeks · 18–22 learning hours  |  Language: English  |  Certificate: Professional Certificate (non-degree)  |  Format: Self-paced with AI support under human supervision

Overview

Executives do not lack information. They lack the time to interrogate it, and they operate in an environment where whoever writes the briefing shapes the decision. A generative system inserted into that environment can improve decision quality by producing the counterargument nobody in the room will make, or degrade it by producing a fluent consensus that no one challenges. Which of these happens is determined by how the tool is used, not by the tool.

This course is short, practical and pitched at decisions rather than technique. It covers three applications with the strongest evidence base: generating alternatives that the executive team has not considered, constructing the case against a proposal the executive already favours, and pre-mortem analysis of a decision before commitment. It also covers the two failure modes with the greatest cost: the disappearance of dissent, and the manufactured confidence of a well-written briefing built on thin evidence.

Learners apply the techniques to a real, live decision in their own organisation and produce a decision record that documents the alternatives considered, the evidence weighed, the confidence held, and the conditions under which the decision should be revisited.

Learning outcomes

On completion, a successful learner will be able to:

1. Identify the decisions in your portfolio where structured analysis would change the outcome, and those where it would not.
2. Use machine assistance to generate genuine alternatives and to construct the strongest case against your preferred option.
3. Conduct a pre-mortem and convert its findings into monitored conditions.
4. Detect and counteract the loss of dissent in a decision process that uses machine-drafted material.
5. Interrogate a briefing for evidential thinness beneath fluent presentation.
6. Produce a decision record that makes reasoning, confidence and revision conditions explicit.

Who this course is for

Rectors, directors, secretaries-general, board members, chief executives and senior officials, together with the chiefs of staff and advisers who prepare their decisions.

Prerequisites

Senior decision-making responsibility. No technical prerequisite. Learners must bring one real, unresolved decision; the assessed work is applied to it and remains confidential to the learner and assessor.

Syllabus

Module 1 — Which decisions are worth the process

Focus. Consequence, reversibility and frequency as the triage criteria. Recognising the decisions where structured analysis pays and the many where it merely delays.

Lessons. 1.1 Consequence, reversibility, frequency. 1.2 Decisions that deserve process. 1.3 Decisions that deserve speed. 1.4 The cost of deliberation.

Core reading. John S. Hammond, Ralph L. Keeney & Howard Raiffa, Smart Choices: A Practical Guide to Making Better Decisions (Boston: Harvard Business School Press, 1999), chapters 1–3. David J. Snowden & Mary E. Boone, “A Leader’s Framework for Decision Making”, Harvard Business Review 85, no. 11 (November 2007).

Deliverable. Decision triage of your current portfolio, with the one live decision selected and justified.

Module 2 — Alternatives you have not considered

Focus. Most executive decisions are made between two options because only two were presented. Systematic alternative generation, and using a model to widen the option set before narrowing it.

Lessons. 2.1 The narrow frame problem. 2.2 Systematic alternative generation. 2.3 Machine assistance for option widening. 2.4 Discarding alternatives with reasons.

Core reading. Hammond, Keeney & Raiffa, Smart Choices, chapters 4–6. Dan Lovallo & Daniel Kahneman, “Delusions of Success: How Optimism Undermines Executives’ Decisions”, Harvard Business Review 81, no. 7 (July 2003): 56–63.

Deliverable. Option set of at least six alternatives, with the reasons for discarding each rejected one.

Module 3 — The case against your own view

Focus. Constructing the strongest available argument against the option you prefer, and doing so with a tool that has no career interest in the answer. Pre-mortem analysis and converting it into monitored conditions.

Lessons. 3.1 Steel-manning the opposing case. 3.2 Pre-mortem method. 3.3 From failure scenario to monitored indicator. 3.4 Deciding what would change your mind.

Core reading. Michael A. Roberto, Why Great Leaders Don’t Take Yes for an Answer, 2nd edition (Upper Saddle River: Pearson, 2013), chapters 3–5. Philip E. Tetlock & Dan Gardner, Superforecasting (New York: Crown, 2015), chapters 5–6.

Deliverable. Adversarial case and pre-mortem, with monitored indicators specified.

Module 4 — Where dissent goes

Focus. How machine-drafted material suppresses disagreement, why a fluent briefing is harder to challenge than a rough one, and process design that keeps dissent alive without theatre.

Lessons. 4.1 Fluency and the suppression of objection. 4.2 Noise, bias and the executive team. 4.3 Process design for real dissent. 4.4 Interrogating a briefing for evidential thinness.

Core reading. Daniel Kahneman, Olivier Sibony & Cass R. Sunstein, Noise: A Flaw in Human Judgment (New York: Little, Brown Spark, 2021), parts III–V. Roberto, Why Great Leaders Don’t Take Yes for an Answer, chapters 6–8.

Deliverable. Process design note for your own decision forum, with three interrogation questions for any briefing.

Module 5 — The decision record

Focus. Recording reasoning, confidence and revision conditions so that the decision can be evaluated later on its process rather than only its outcome. Accountability for machine-assisted decisions.

Lessons. 5.1 What a decision record contains. 5.2 Confidence and its later scoring. 5.3 Revision conditions and sunk cost. 5.4 Accountability that cannot be delegated to a system.

Core reading. Thomas H. Davenport & Rajeev Ronanki, “Artificial Intelligence for the Real World”, Harvard Business Review 96, no. 1 (January–February 2018). Kahneman, Sibony & Sunstein, Noise, part VI.

Deliverable. Final submission: complete decision record for the live decision, with confidence, monitored indicators and revision conditions.

Assessment

Pass mark 70 per cent. All assessed components must be attempted. Every mark in this course is issued by a human assessor; no assessment outcome is generated automatically.

Rubric criteria

Each assessed artefact is marked against four criteria at four levels (distinction, pass with merit, pass, fail).

1. Genuine widening: are the additional alternatives real options rather than variations of the preferred one?
2. Strength of the adversarial case: would a competent opponent recognise their own best argument here?
3. Process realism: could the dissent design be run in the learner’s actual forum, with its actual politics?
4. Accountability: does the decision record leave a named human accountable, with revision conditions that could trigger?

Reading list

Core. John S. Hammond, Ralph L. Keeney & Howard Raiffa, Smart Choices (Boston: Harvard Business School Press, 1999). Michael A. Roberto, Why Great Leaders Don’t Take Yes for an Answer, 2nd edition (Upper Saddle River: Pearson, 2013). Daniel Kahneman, Olivier Sibony & Cass R. Sunstein, Noise: A Flaw in Human Judgment (New York: Little, Brown Spark, 2021).

Articles. David J. Snowden & Mary E. Boone, “A Leader’s Framework for Decision Making”, Harvard Business Review (November 2007). Dan Lovallo & Daniel Kahneman, “Delusions of Success”, Harvard Business Review (July 2003). Thomas H. Davenport & Rajeev Ronanki, “Artificial Intelligence for the Real World”, Harvard Business Review (January–February 2018).

Wider reading. Philip E. Tetlock & Dan Gardner, Superforecasting (New York: Crown, 2015). Daniel Kahneman, Thinking, Fast and Slow (New York: Farrar, Straus and Giroux, 2011).

All items are published works identifiable by author, title and publisher. Learners obtain them through an institutional library or the publisher. The Academy does not distribute copyrighted texts.

Academic integrity and use of AI

Generative tools may be used in producing assessed work under three conditions. Use must be disclosed in a short statement appended to each submission, naming the tool and the task it performed. Any factual or technical claim originating from a generative tool must be verified against a citable source before it enters assessed work, and the verification must be evidenced. The analytical judgement in each artefact must be the learner’s own and must be defensible in a short follow-up. Decisions submitted for assessment are treated as confidential to the learner and the assessor. Learners must not submit material they are not authorised to share outside their organisation.

Instructor: pending owner confirmation. Pricing: pending owner approval. Reference list verified against publisher records; any later addition is marked for verification before publication.