Abstract 3

Acknowledgements 4

Abbreviations and acronyms 7

1 Executive summary 8

AI in public audit: An emergent but rapidly evolving landscape 8

Infrastructure and capacity: Bridging the gap between pilots and production 8

Data governance: The foundation that AI cannot afford to lack 8

Shared challenges and emerging good practices 9

2 Overview and state of play 10

3 Use cases 14

Anomaly detection and risk assessments 14

AI for classification in audit and oversight 16

Predictive and preventive audit models 18

Search, retrieval and knowledge management 19

Intelligent Document Processing (IDP) and data extraction 21

Generative AI for drafting, summarising and translating 23

Visual and spatial data analysis 25

4 Infrastructure and capacity 30

Hardware infrastructure 30

Software environment and stack 31

Data infrastructure and governance 32

Development capabilities 34

5 Measuring impact 36

6 Challenges and strategic considerations 38

Ethics, explainability and bias 38

Organisational culture, capacity and supporting adoption 39

Data quality as a strategic asset 40

Governance and security 40

Managing the pace of change and reducing third-party dependencies 40

Balancing innovation, long-term stability and sustainability 41

7 Good practices and success factors 44

Transparency and explainability 44

User and stakeholder involvement: facilitating bottom-up innovation 44

Management and leadership support 45

Institutional governance and oversight 47

Effective training and knowledge sharing 47

Strategic partnerships 48

8 Future directions 50

9 Conclusions 52

Glossary 54

References 57