AI Automation Potential
Which government agencies have the highest potential for AI-driven efficiency gains? Score based on 5 proxy metrics from annual report data.
55
Avg AI Potential Score
6
High/Very High Agencies
$2.6B
Est. Addressable Spend
21
Agencies Analysed
Est. addressable: $47M
50
75
90
50
75
Est. addressable: $5M
80
75
75
50
30
Est. addressable: $229M
60
75
90
50
30
Est. addressable: $1913M
60
60
75
50
60
Est. addressable: $100M
60
60
90
50
45
Est. addressable: $4M
50
90
75
50
30
Est. addressable: $29M
60
60
75
50
45
Est. addressable: $191M
45
75
75
50
30
Est. addressable: $4M
45
75
75
50
30
Est. addressable: $17M
60
50
50
50
75
50
50
60
50
60
50
50
75
50
45
Est. addressable: $22M
50
30
75
50
60
50
50
75
50
30
Est. addressable: $26M
30
45
75
50
60
Est. addressable: $6M
30
50
75
50
30
Est. addressable: $8M
50
30
75
50
30
Est. addressable: $69K
30
50
75
50
30
50
50
50
50
30
Est. addressable: $1M
50
30
60
50
30
Est. addressable: $33M
50
30
45
50
45
Methodology (v1.0)
The AI Automation Potential Score estimates how much of an agency's current operations could be augmented or automated with AI/ML technologies, based on observable proxy metrics from annual reports.
High outsourced spend signals work automatable via AI (analysis, reporting, data processing). McKinsey (2023): "30% of outsourced professional services are AI candidates."
Higher personnel % of total expenses = more labour-intensive operations. OECD (2024): back-office functions show 40-60% automation potential.
FTE per $100M revenue. More staff per dollar = more human processes that AI could augment.
Transactional agencies (IR, MSD, ACC) score higher than policy-only (Treasury, MFAT). Based on NZ Digital Strategy + Deloitte Gov Trends 2024.
Count of "AI", "automation", "digital transformation", "machine learning" mentions in annual report. More mentions = more identified use cases.
Estimated Addressable Spend = (AI score % x contractor spend) + (15% x personnel costs x AI score %). This is a rough order-of-magnitude estimate, not a savings projection.
Important: This score indicates automation POTENTIAL, not a recommendation. Not all work should be automated — policy judgment, human empathy, complex negotiations, and democratic accountability require human involvement. The score helps identify where AI could free up human capacity for higher-value work.
References: McKinsey Global Institute (2023) "The economic potential of generative AI"; OECD (2024) "AI in the Public Sector"; NZ Gov (2022) "Digital Strategy for Aotearoa"; Deloitte (2023) "Government Trends 2024".
Data sourced from publicly available NZX filings and annual reports. Our datasets may not be complete. Automated analysis can produce errors. Scores are calculated using disclosed methodology and are analytical tools, not investment ratings or recommendations. If you believe any data on this page is incorrect, please contact us at hello@nzxplorer.co.nz. For informational purposes only. Not investment advice.