AI-G02 — AI Readiness Assessment and Institutional Roadmapping

Course Code: AI-G02  |  School: School of Artificial Intelligence  |  Cluster: Level 4 — Governance & Policy

Level: Advanced  |  Duration: 6 weeks · 24–30 learning hours  |  Language: English  |  Certificate: Executive Certificate (non-degree)  |  Format: Self-paced with AI support under human supervision

Overview

Readiness is not enthusiasm, and it is not a technology audit. An institution is ready for AI when it can name the decisions it wants to improve, has data it can actually locate and trust, has staff who can supervise the output, and has a governance route for approving use. This course teaches learners to assess those four things honestly and to build a roadmap that survives contact with a budget cycle.

The assessment instrument is built during the course, not handed over ready-made, because a borrowed maturity model produces borrowed conclusions. Learners define dimensions, evidence requirements and scoring rules, then run the assessment on their own institution and defend the result to a sceptical reader.

The roadmap that follows is deliberately conservative in sequence: capability before ambition, one or two decisions before a platform, and a named owner before a pilot. The most common institutional failure this course is designed to prevent is a signed enterprise agreement with no internal capacity to use it.

Learning outcomes

On completion, a successful learner will be able to:

1. Define the decisions or processes an AI investment is supposed to improve, in terms that make success measurable.
2. Design and apply a readiness assessment across strategy, data, capability, governance and infrastructure, with evidence requirements for each score.
3. Locate, describe and appraise the institutional data that a proposed use would depend on, including its provenance and permitted use.
4. Assess workforce capability honestly, distinguishing tool familiarity from supervisory competence.
5. Sequence a two-to-three-year roadmap with dependencies, decision gates and stop conditions.
6. Build a cost and benefit case that includes maintenance, capability building and the cost of stopping.
7. Present findings to a governing body, including the finding that the institution is not ready.

Who this course is for

Institutional leaders and planners: vice-rectors, directors of planning, heads of digital transformation, secretaries of ministries and agencies, and consultants conducting readiness reviews.

Prerequisites

AI-F01 or equivalent. Familiarity with your institution’s planning and budget cycle. Learners need access to institutional documents and to at least three colleagues who can be interviewed for the assessed work.

Syllabus

Module 1 — Starting from the decision, not the technology

Focus. Identifying the decisions and processes worth improving, and writing success criteria that can be evaluated later. Why capability-led adoption produces orphaned tools.

Lessons. 1.1 Decisions, processes and outcomes. 1.2 Writing success criteria that can be tested. 1.3 The orphaned-tool pattern. 1.4 Framing the assessment question.

Core reading. Thomas H. Davenport & Rajeev Ronanki, “Artificial Intelligence for the Real World”, Harvard Business Review 96, no. 1 (January–February 2018): 108–116.

Deliverable. Decision inventory: five candidate decisions or processes with measurable success criteria.

Module 2 — Designing the readiness instrument

Focus. Constructing dimensions, levels and evidence requirements. Why a score without a required evidence type is a survey of optimism rather than an assessment.

Lessons. 2.1 Dimensions that matter: strategy, data, capability, governance, infrastructure. 2.2 Levels and observable indicators. 2.3 Evidence requirements per level. 2.4 Avoiding self-report inflation.

Core reading. NIST, AI Risk Management Framework (AI RMF 1.0), NIST AI 100-1 (2023), GOVERN and MAP functions, used as a source of dimension structure.

Deliverable. Readiness instrument with dimensions, levels and evidence requirements.

Module 3 — Data reality

Focus. Where institutional data actually lives, who owns it, whether it may lawfully be used for the intended purpose, and what condition it is in. Provenance, quality and permitted use as three separate questions.

Lessons. 3.1 Locating and describing holdings. 3.2 Provenance and lawful basis. 3.3 Quality, completeness and consistency. 3.4 What must be fixed before any pilot.

Core reading. Timnit Gebru et al., “Datasheets for Datasets”, Communications of the ACM 64, no. 12 (2021). Undang-Undang No. 27 Tahun 2022 tentang Pelindungan Data Pribadi (Indonesia), or the applicable data protection instrument in the learner’s jurisdiction.

Deliverable. Data appraisal for the two highest-priority candidate uses.

Module 4 — Capability and the supervision gap

Focus. Distinguishing tool familiarity from the ability to supervise machine output. Estimating the training required, and identifying the roles that must exist before deployment.

Lessons. 4.1 Familiarity versus supervisory competence. 4.2 Role inventory: who must be able to do what. 4.3 Training estimates and realistic timescales. 4.4 Retention and single points of failure.

Core reading. Erik Brynjolfsson, Daniel Rock & Chad Syverson, “Artificial Intelligence and the Modern Productivity Paradox”, NBER Working Paper 24001 (National Bureau of Economic Research, 2017).

Deliverable. Capability assessment with role inventory and training estimate.

Module 5 — Roadmap, gates and stop conditions

Focus. Sequencing by dependency rather than ambition. Decision gates, pilot design with pre-registered success criteria, and explicit stop conditions so that failure can be recognised.

Lessons. 5.1 Dependency-first sequencing. 5.2 Decision gates and what passes them. 5.3 Pilot design and pre-registered criteria. 5.4 Stop conditions and sunk-cost discipline.

Core reading. Tim Fountaine, Brian McCarthy & Tamim Saleh, “Building the AI-Powered Organization”, Harvard Business Review 97, no. 4 (July–August 2019): 62–73.

Deliverable. Roadmap with dependencies, gates and stop conditions.

Module 6 — The case, and the courage to report unreadiness

Focus. Cost and benefit including maintenance and capability building, the cost of stopping, and presenting an honest finding to a governing body that may have already announced an ambition.

Lessons. 6.1 Total cost including maintenance and drift. 6.2 Benefit estimation without inflation. 6.3 Reporting unreadiness constructively. 6.4 The board presentation.

Core reading. George Westerman, Didier Bonnet & Andrew McAfee, Leading Digital: Turning Technology into Business Transformation (Boston: Harvard Business Review Press, 2014), chapters 1–3.

Deliverable. Final submission: readiness report, roadmap and board-ready presentation.

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. Evidential honesty: are scores supported by documents, interviews or observation rather than by self-report?
2. Diagnostic accuracy: does the assessment identify the binding constraint rather than the most visible one?
3. Sequencing logic: does the roadmap respect dependencies, and are stop conditions specified before commitment?
4. Executive usability: could a governing body make a funding decision from this report without further analysis?

Reading list

Core. George Westerman, Didier Bonnet & Andrew McAfee, Leading Digital: Turning Technology into Business Transformation (Boston: Harvard Business Review Press, 2014).

Articles. Thomas H. Davenport & Rajeev Ronanki, “Artificial Intelligence for the Real World”, Harvard Business Review (January–February 2018). Tim Fountaine, Brian McCarthy & Tamim Saleh, “Building the AI-Powered Organization”, Harvard Business Review (July–August 2019).

Working paper. Erik Brynjolfsson, Daniel Rock & Chad Syverson, “Artificial Intelligence and the Modern Productivity Paradox”, NBER Working Paper 24001 (2017).

Peer-reviewed and standards. Timnit Gebru et al., “Datasheets for Datasets”, Communications of the ACM 64, no. 12 (2021). NIST, AI Risk Management Framework (AI RMF 1.0) (2023). ISO/IEC 42001:2023.

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. A readiness report whose conclusion was decided before the evidence was gathered fails the course regardless of presentation quality.

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