Generated 2026-05-31

NFP AI in Australia

What Australian not-for-profits say about AI in their annual reports.

5,177 Representative reports

from 1,628 coded NFP entities

1.1% AI-positive estimate

57/5,177 representative reports

21,369 Evidence corpus

344 AI-positive reports from 5,845 entities

87 Own-use evidence

143 documents across 87 distinct entities

Executive Read

What the evidence currently supports

The reports do not show an NFP sector already transformed by AI. They show early, uneven use: mostly internal work, service delivery, and governance starting to appear.

  • AI appears in a small share of representative reports: 57/5,177 are AI-positive (1.1%).
  • The full deduped evidence corpus contains 344 AI-positive reports, including 143 own-use documents across 87 entities.
  • Most disclosed use is internal capability, service delivery, program delivery, and administrative work, not a wave of external AI products.
  • Research and commentary matter: 101 AI-positive reports are about research subject matter and 46 are sector commentary rather than own organisational adoption.
  • Governance is uneven: 70 AI-positive reports mention formal or informal governance, while 53 describe AI in regular use.
  • AI is more often framed as opportunity than risk (150 opportunity-led reports vs 87 balanced or risk-led reports), and named barriers are rare.
Chapter 1

How much AI appears in annual reports?

The first question is not whether AI is everywhere; it is what can be said from annual reports alone. In the representative sample, AI disclosure is still sparse: 57 of 5,177 coded reports are AI-positive. Across the full deduped evidence corpus, there are 344 AI-positive reports, including 143 documents showing the organisation's own use.

  • The representative sample covers 5,000 ACNC NFP entities and 5,177 coded annual reports.
  • The largest-500 slice mentions AI more often than the representative sample (5.5% vs 1.1%).
  • The revenue-ranked comparison scan sits at 1.8% AI-positive across 17,824 coded reports.
  • Year-to-year movement should be read with report coverage in mind rather than as a complete sector adoption curve.

Prevalence by Cohort

AI-positive share by cohort; dots mark reports coded as own use.

Trend by Financial Year

Sorted by financial year; detailed denominators are in the appendix.
Chapter 2

Where do reports mention AI most?

Evidence-location charts use the full deduped coded corpus. Representative prevalence charts are kept as cautious rate estimates where the denominator is strong enough. AI evidence is unevenly distributed. It clusters around larger organisations and organisations already doing research, data work, education, health, or complex service delivery, while representative prevalence estimates remain sparse in many cells.

  • Education carries much of the disclosed AI-positive evidence, with health, general public benefit, social welfare, culture, and environment providing additional examples.
  • State differences should be read as disclosure patterns, not league tables; they mix sector composition, report availability, and organisational scale.
  • Representative sector and state prevalence charts estimate rates, while the full-corpus evidence charts show where the richest examples are.

AI-Positive Evidence by Sector

Full deduped coded NFP corpus; counts of AI-positive reports, not population prevalence.

AI-Positive Evidence by State

Full deduped coded NFP corpus; counts of AI-positive reports, not population prevalence.

Sector Prevalence

Representative 5,000 only; sectors with at least 100 coded reports.

State Prevalence

Representative 5,000 only; state and territory cells with at least 30 coded reports.

Size Category

Representative 5,000 only. Small and medium annual-report cells are thin, so this should not be read as evidence of no operational AI use.
Chapter 3

What are the AI mentions actually about?

This chapter uses the full deduped AI-positive evidence corpus. Not every AI mention means the organisation is adopting AI. This distinction matters: not every AI mention means the organisation is using AI itself. Of the 344 AI-positive reports in the full evidence corpus, 143 are about the organisation's own adoption, while 101 are research subject matter, 46 are sector commentary, and 13 are risk or governance discussion without own-use evidence.

  • A mention of AI is not the same thing as adopting AI.
  • Research-heavy organisations create many legitimate AI mentions that should not be counted as operational adoption.
  • Governance-only mentions matter because they show board or risk attention before a concrete use case appears.

What AI Mentions Are About

Full deduped AI-positive corpus; separates own use from research, commentary, governance-only discussion, and false-positive mentions.
Chapter 4

What does own-use adoption look like?

For reports with own-use evidence, the useful story is what the AI is being used for: maturity, setting, technology type, and concrete use cases. This chapter uses the full deduped own-use evidence corpus. Where organisations describe their own use, it is usually practical. The reports mostly describe internal service delivery, internal operations, program delivery, student or client support, and data analysis rather than organisations selling AI products to the public.

  • Internal service delivery appears in 46 own-use documents and internal operations in 32; external-product evidence appears in 13.
  • Some reports describe AI already in regular use (53 AI-positive reports), but embedded practice remains uncommon.
  • Client and service delivery (53 docs); Program and education delivery (52 docs); Admin and productivity (43 docs)

AI Practice Maturity

Full deduped own-use corpus; AI practice maturity among reports with own-use evidence.

Usage Orientation

Full deduped own-use corpus; shows how reports divide across internal service delivery, operations, research, external products, and mixed uses.

AI Types

Full deduped AI-positive corpus; unclear remains high where annual reports are nonspecific.

Use-Case Families

Full deduped own-use corpus; detailed use-case labels are grouped into simpler families.

Use Cases by Orientation

Full deduped own-use corpus; main use-case families split by internal, research, external, or mixed orientation.
Chapter 5

Governance, framing, and constraints

Governance does not always move at the same pace as use. This chapter uses the full deduped AI-positive and own-use evidence corpus, so it describes disclosed patterns rather than population prevalence. The governance story is not simply maturity catching up with adoption. Some organisations are creating policies, steering groups, staff guidance, and risk discussions before regular use is described, while others report regular use without mentioning governance.

  • At entity level, 16 own-use entities mention governance without regular-use evidence, while 18 describe regular use without mentioning governance.
  • Formal or informal governance appears in 70 AI-positive reports; 53 reports describe AI in regular use.
  • AI is framed as opportunity-led in 150 reports, while balanced or risk-led framing appears in 87. Capability is the most common named enabler (78 reports), and 334 reports name no concrete barrier.

AI Governance Mentioned

Full deduped AI-positive corpus; governance is coded separately from the organisation's own AI use.

Governance and Routine Use

Full deduped own-use corpus; some organisations mention governance before they describe regular AI use.

Governance by Maturity

Full deduped own-use corpus; reports grouped by practice maturity and whether governance is mentioned.

AI Framing

Full deduped AI-positive corpus; opportunity, risk, balanced, or neutral framing.

Investment Intent

Full deduped AI-positive corpus; whether reports mention increasing, maintaining, or no AI investment intent.

Primary Enablers

Full deduped AI-positive corpus; the main enabler coded in each report.

Primary Barriers

Full deduped AI-positive corpus; most reports do not say what is stopping or slowing AI use.
Appendix

Data table appendix

Detailed tables are kept here as the audit trail behind the figures. They are closed by default so the main report remains chart-led.

Table · Chapter 1 Coverage and Denominators 3 rows

The cohort exists at entity level; the coded denominator is the subset with annual-report text and completed NFP-AI classifications.

Cohort Universe entities Coded entities Coverage % Reports AI-positive reports AI-positive % Own-use reports
Representative 5,000 5000 1628 32.56 5177 57 1.1 16
Revenue-ranked 5,181 5181 4619 89.15 17824 312 1.75 136
Largest 500 500 493 98.6 2444 134 5.48 75
Table · Chapter 1 Financial-Year Trend 18 rows

This is the audit trail behind the year-by-year movement. The final-year denominator is smaller, so recent-year comparisons should be treated cautiously.

Cohort FY Reports AI-positive AI-positive % Own use
Largest 500 FY2020 385 18 4.68 8
Largest 500 FY2021 426 21 4.93 11
Largest 500 FY2022 438 25 5.71 12
Largest 500 FY2023 450 26 5.78 17
Largest 500 FY2024 446 28 6.28 18
Largest 500 FY2025 299 16 5.35 9
Representative 5,000 FY2020 718 5 0.7 0
Representative 5,000 FY2021 821 5 0.61 1
Representative 5,000 FY2022 869 6 0.69 1
Representative 5,000 FY2023 973 11 1.13 2
Representative 5,000 FY2024 979 16 1.63 6
Representative 5,000 FY2025 817 14 1.71 6
Revenue-ranked 5,181 FY2020 2711 37 1.36 14
Revenue-ranked 5,181 FY2021 2990 42 1.4 17
Revenue-ranked 5,181 FY2022 3121 52 1.67 19
Revenue-ranked 5,181 FY2023 3224 59 1.83 26
Revenue-ranked 5,181 FY2024 3263 71 2.18 31
Revenue-ranked 5,181 FY2025 2515 51 2.03 29
Table · Chapter 2 AI-Positive Evidence by Sector 9 rows

Counts use the full deduped coded NFP corpus and show where AI-positive evidence appears. They are not population prevalence estimates.

Sector AI-positive reports AI-positive entities Own-use reports Own-use entities % AI-positive evidence
Education 139 66 59 31 40.41
Health 64 39 28 18 18.6
General Public Benefit 29 14 11 8 8.43
Social Welfare 21 16 11 8 6.1
Culture 16 12 4 3 4.65
Environment 13 5 1 1 3.78
Advocacy 4 1 4 1 1.16
Religion 4 2 0 0 1.16
Security & Safety 4 2 1 1 1.16
Table · Chapter 2 AI-Positive Evidence by State 7 rows

Counts use the full deduped coded NFP corpus and show where AI-positive evidence appears. They are not state league tables.

State AI-positive reports AI-positive entities Own-use reports Own-use entities % AI-positive evidence
NSW 131 67 58 31 38.08
VIC 106 57 41 25 30.81
WA 33 19 12 10 9.59
QLD 23 17 15 11 6.69
ACT 22 9 8 3 6.4
SA 10 5 3 3 2.91
TAS 6 3 1 1 1.74
Table · Chapter 2 State and Territory Breakdown 27 rows

State cells use the same report-level denominator as the prevalence charts. Thin cells and missing-state rows should be read cautiously.

Cohort State Reports AI-positive AI-positive % Own use Own-use %
Largest 500 NSW 799 56 7.01 29 3.63
Largest 500 VIC 680 42 6.18 25 3.68
Largest 500 QLD 322 10 3.11 8 2.48
Largest 500 WA 236 13 5.51 6 2.54
Largest 500 SA 180 1 0.56 1 0.56
Largest 500 (missing) 85 4 4.71 0 0
Largest 500 ACT 61 8 13.11 6 9.84
Largest 500 NT 44 0 0 0 0
Largest 500 TAS 37 0 0 0 0
Representative 5,000 NSW 1503 14 0.93 7 0.47
Representative 5,000 VIC 1320 20 1.52 2 0.15
Representative 5,000 QLD 776 2 0.26 2 0.26
Representative 5,000 WA 473 9 1.9 3 0.63
Representative 5,000 (missing) 434 1 0.23 0 0
Representative 5,000 SA 332 6 1.81 1 0.3
Representative 5,000 ACT 119 1 0.84 0 0
Representative 5,000 NT 113 0 0 0 0
Representative 5,000 TAS 107 4 3.74 1 0.93
Revenue-ranked 5,181 NSW 5423 121 2.23 54 1
Revenue-ranked 5,181 VIC 4707 98 2.08 41 0.87
Revenue-ranked 5,181 QLD 2661 22 0.83 14 0.53
Revenue-ranked 5,181 WA 1654 31 1.87 12 0.73
Revenue-ranked 5,181 SA 1129 4 0.35 2 0.18
Revenue-ranked 5,181 (missing) 896 13 1.45 5 0.56
Revenue-ranked 5,181 ACT 554 21 3.79 8 1.44
Revenue-ranked 5,181 TAS 410 2 0.49 0 0
Revenue-ranked 5,181 NT 390 0 0 0 0
Table · Chapter 3 What AI Mentions Are About 6 rows

Full deduped AI-positive corpus. This gate separates own organisational adoption from AI as a research subject, sector commentary, risk/governance-only discussion, or false-positive AI mentions.

Mention is about Reports %
own_adoption 143 41.57
research_subject 101 29.36
sector_commentary 46 13.37
not_about_ai 40 11.63
risk_or_governance_only 13 3.78
unclear 1 0.29
Table · Chapter 4 AI Practice Maturity 7 rows

Full deduped own-use corpus. Practice-maturity categories are derived from the coded adoption-stage field and remain provisional until blind human IRR certifies the field.

AI practice maturity Reports %
Absent 5 3.5
Assessing opportunities 40 27.97
Governed intent 27 18.88
Limited trial 14 9.79
Routine use 49 34.27
Embedded practice 5 3.5
Multiple stages 3 2.1
Table · Chapter 4 Usage Orientation 7 rows
Usage orientation Reports %
internal_service_delivery 46 32.17
multiple 38 26.57
internal_ops 32 22.38
external_product 13 9.09
research_rd 12 8.39
none 1 0.7
unclear 1 0.7
Table · Chapter 4 AI Types 8 rows
AI type Reports % AI-positive
unclear 149 43.31
ml 94 27.33
other 83 24.13
genai 46 13.37
none 41 11.92
automation 24 6.98
robotics 14 4.07
cv 10 2.91
Table · Chapter 4 Use-Case Families 7 rows

Detailed project labels grouped into simpler families for the main report.

Use-case family Projects Reports Entities % own docs Orientation Example
Client and service delivery 81 53 40 37.06 external_product|internal_ops|internal_service_delivery|multiple|research_rd Australian Red Cross Society
Program and education delivery 94 52 36 36.36 external_product|internal_ops|internal_service_delivery|multiple Australian Red Cross Society
Admin and productivity 63 43 38 30.07 internal_ops|multiple|unclear Griffith University
Other or unclear 56 34 25 23.78 external_product|internal_ops|internal_service_delivery|multiple|research_rd|unclear Australian Medical Council Limited
Research, data and prediction 57 28 22 19.58 external_product|internal_ops|internal_service_delivery|multiple|research_rd Spatial Information Systems Research Ltd
External tools and products 16 11 9 7.69 external_product|research_rd Griffith University
Cybersecurity and safety 7 7 7 4.9 external_product|internal_ops|research_rd Royal Life Saving Society Of Australia (New South Wales Branch)
Table · Chapter 4 Use-Case Families by Orientation 25 rows

Shows whether top use-case families are internal, external, research-oriented, or mixed.

Use-case family Orientation Projects Reports
Client and service delivery Internal service delivery 64 41
Client and service delivery Internal operations 2 2
Client and service delivery Research and R&D 3 3
Client and service delivery External product 11 9
Client and service delivery Multiple orientations 1 1
Program and education delivery Internal service delivery 59 36
Program and education delivery Internal operations 10 7
Program and education delivery External product 8 7
Program and education delivery Multiple orientations 17 14
Admin and productivity Internal operations 60 41
Admin and productivity Multiple orientations 2 2
Admin and productivity Unclear 1 1
Research, data and prediction Internal service delivery 5 3
Research, data and prediction Internal operations 5 5
Research, data and prediction Research and R&D 36 15
Research, data and prediction External product 8 6
Research, data and prediction Multiple orientations 3 2
External tools and products Research and R&D 5 4
External tools and products External product 11 9
Other or unclear Internal service delivery 9 8
Other or unclear Internal operations 17 12
Other or unclear Research and R&D 15 8
Other or unclear External product 10 7
Other or unclear Multiple orientations 4 4
Other or unclear Unclear 1 1
Table · Chapter 4 Detailed Use-Case Labels 95 rows

Detailed extracted use-case labels remain here as audit material. The main chart groups them into simpler families.

Use case Projects Reports Entities % own docs Orientation Example
client_service 54 37 30 25.87 external_product|internal_service_delivery Australian Red Cross Society
program_delivery 49 37 30 25.87 external_product|internal_service_delivery|multiple Australian Red Cross Society
admin 49 33 32 23.08 internal_ops|multiple|unclear Griffith University
research 27 12 11 8.39 research_rd Griffith University
education 11 9 9 6.29 external_product|internal_ops|internal_service_delivery Griffith University
product_development 11 7 7 4.9 external_product|research_rd Griffith University
teaching_learning 8 6 6 4.2 external_product|internal_service_delivery Queensland University Of Technology
training 8 6 6 4.2 internal_ops|internal_service_delivery|multiple Independent Living Assessment Incorporated
productivity 5 5 5 3.5 internal_ops|multiple The University of Sydney
data_analysis 5 4 3 2.8 external_product|internal_ops|research_rd Queensland University Of Technology
consulting 4 4 3 2.8 external_product Bush Heritage Australia
decision_support 4 4 3 2.8 internal_service_delivery|research_rd The Health Roundtable Limited
student_support 9 3 3 2.1 internal_service_delivery Monash University
curriculum 7 3 2 2.1 internal_ops|internal_service_delivery|multiple Swinburne University Of Technology
data_reporting 4 3 2 2.1 external_product|internal_ops|internal_service_delivery Bush Heritage Australia
prediction_targeting 4 3 3 2.1 external_product|internal_service_delivery Bush Heritage Australia
chatbot_assistant 3 3 3 2.1 internal_ops|internal_service_delivery Legal Aid Commission Of NSW
communications 3 3 3 2.1 external_product Australian Red Cross Society
customer_service 3 3 3 2.1 external_product|internal_ops|multiple Legal Aid Commission Of NSW
security 3 3 3 2.1 external_product|internal_ops Federation University Australia
rd 4 2 2 1.4 research_rd Monash University
diagnostic 3 2 2 1.4 external_product|internal_service_delivery Lions Eye Institute Limited
health 3 2 2 1.4 external_product|research_rd Swinburne University Of Technology
strategy_governance 3 2 2 1.4 internal_ops|multiple The University Of Wollongong
back_office 2 2 2 1.4 internal_ops Ability Works Australia Ltd
data_management 2 2 1 1.4 internal_ops Melanoma Institute Australia
governance 2 2 2 1.4 internal_ops The University of Sydney
other 2 2 2 1.4 external_product|unclear Australian National University
research_output 2 2 2 1.4 multiple|research_rd The University of Sydney
risk_management 2 2 2 1.4 internal_ops Royal Life Saving Society Of Australia (New South Wales Branch)
Table · Chapter 4 Own-Use Entities 87 rows

The first forty entities with own-use evidence, ordered by routine-use evidence, practice maturity, and entity name.

Entity Cohort First FY Last FY Sector Max stage Orientation Routine use? Governance Use cases
Sachi Foundation Representative 5,000 2025 2025 Education Multiple stages multiple True formal admin|client_service|program_delivery|research
The University of Sydney Largest 500 2023 2024 Education Multiple stages multiple True formal admin|education|governance|productivity|program_delivery|research|research_output|strategy|teaching_learning|training_development
Australian National University Largest 500 2020 2023 Education Embedded practice external_product|multiple True formal curriculum|education|health|other|partnerships|program_delivery|public_policy|research
Macquarie University Largest 500 2021 2023 Education Embedded practice internal_service_delivery|multiple True formal admin|client_service|education|process_improvement|product_development|program_delivery|public_good|research|security|student_support|teaching_support
Swinburne University Of Technology Largest 500 2021 2024 Education Embedded practice multiple True formal admin|communications|consulting|curriculum|health|manufacturing|optimisation|product_development|program_delivery|research|training
Australian Red Cross Society Largest 500 2023 2024 Routine use multiple True formal client_service|communications|program_delivery
Butterfly Foundation Revenue-ranked 5,181 2021 2021 Health Routine use external_product True none client_service
Child Abuse Prevention Service (Sydney) Inc Representative 5,000 2021 2024 Social Welfare Routine use external_product|internal_service_delivery True none child_protection_risk_assessment|client_service|program_delivery
Cotton Seed Distributors Ltd Largest 500 2020 2024 General Public Benefit Routine use internal_service_delivery|research_rd True none client_service|data_reporting|prediction_targeting|product_development
Deakin University Largest 500 2020 2024 Education Routine use multiple True formal back_office_ops|client_service|program_delivery|research_development|strategy_governance|training_knowledge_management
Gordon Institute Of Tafe Largest 500 2022 2023 Education Routine use internal_service_delivery|research_rd True none workforce_development
Intouch Multicultural Centre Against Family Violence Revenue-ranked 5,181 2024 2025 Culture Routine use internal_service_delivery True none client_service
Islamic College Of Brisbane Limited Revenue-ranked 5,181 2022 2022 Education Routine use multiple True informal admin|client_service|program_delivery
Justice Connect Revenue-ranked 5,181 2023 2023 Social Welfare Routine use internal_service_delivery True none client_service
Law Institute of Victoria Limited Revenue-ranked 5,181 2023 2025 Education Routine use internal_ops|multiple True informal admin|advisory/policy|advisory_advocacy|communications_outreach|education_training
Legal Aid Commission Of NSW Largest 500 2023 2025 Routine use external_product|internal_ops|internal_service_delivery True informal admin|back_office|chatbot_assistant|client_service|customer_service|legal_research|productivity|risk_management
Legal Aid Commission Of Wa Largest 500 2020 2023 Routine use external_product|internal_service_delivery True formal admin|client_service
Legal Services Commission of SA Largest 500 2020 2020 Routine use external_product True none client_service|program_delivery
Lions Eye Institute Limited Representative 5,000 2023 2023 Health Routine use multiple True formal diagnosis|diagnostic|screening
MINEX CRC LTD Revenue-ranked 5,181 2023 2023 Education Routine use research_rd True none data_analysis|research
MTP-IIGC LTD Revenue-ranked 5,181 2021 2021 General Public Benefit Routine use internal_service_delivery True none diagnostic|patient_monitoring
Monash University Largest 500 2020 2024 Education Routine use internal_service_delivery|multiple|research_rd True formal admin|compliance|education|product_development|program_delivery|public_health|public_safety|rd|scientific_research|student_support|teaching_learning|training
NSW Business Chamber Limited Largest 500 2021 2022 General Public Benefit Routine use internal_service_delivery True none client_service|program_delivery
Opportunity International Australia Limited Revenue-ranked 5,181 2025 2025 Social Welfare Routine use internal_service_delivery True none program_delivery
Positive Media Limited Revenue-ranked 5,181 2024 2024 Culture Routine use internal_ops True none content_creation
Queensland University Of Technology Largest 500 2022 2024 Education Routine use multiple True formal client_service|data_analysis|education_delivery|hr/staff_productivity|other_research|program_delivery|research|research_admin|teaching_learning
Reachout Australia Representative 5,000 2024 2025 Health Routine use internal_service_delivery True informal client_service
Southcare Inc Revenue-ranked 5,181 2022 2023 Routine use internal_ops|internal_service_delivery True none client_service|customer_service
Spatial Information Systems Research Ltd Revenue-ranked 5,181 2025 2025 General Public Benefit Routine use external_product True none data_analytics
Speld Qld Inc. Revenue-ranked 5,181 2025 2025 General Public Benefit Routine use internal_service_delivery True none client_service
St. Vincent's Hospital (Melbourne) Limited Largest 500 2020 2025 Health Routine use internal_ops|internal_service_delivery|multiple True formal client_service|clinical_diagnosis|diagnostic_imaging|patient_identification|prediction_targeting|program_delivery|strategy
The Health Roundtable Limited Revenue-ranked 5,181 2020 2025 Health Routine use internal_ops|internal_service_delivery True none data_analysis|decision_support|patient_safety
The University Of Wollongong Largest 500 2020 2024 Education Routine use internal_service_delivery|multiple|research_rd True formal assessment|curriculum_design|curriculum_development|education|policy_compliance|policy_governance|product_development|program_delivery|strategy_governance|student_support
Toowoomba Hospital Foundation Revenue-ranked 5,181 2023 2023 Routine use internal_ops True none marketing
University Of Canberra Largest 500 2021 2023 Education Routine use internal_service_delivery|multiple True formal education|governance_compliance|strategic_planning|teaching_and_learning|teaching_support|training
University Of The Sunshine Coast Largest 500 2021 2024 Education Routine use internal_service_delivery|multiple|research_rd True formal admin|decision_support|governance|program_delivery|teaching_learning|training
Victoria University Representative 5,000 2024 2024 Education Routine use multiple True formal admin|chatbot_assistant|client_service|clinical_care|education|program_delivery|research
Volunteering Queensland Inc Revenue-ranked 5,181 2025 2025 General Public Benefit Routine use internal_ops True none admin|comms_marketing
Griffith University Largest 500 2025 2025 Education Multiple stages multiple False formal admin|education|product_development|research
Black Dog Institute Largest 500 2021 2021 Routine use research_rd False none client_service|diagnosis_treatment|early_warning
Table · Chapter 5 AI Governance Mentioned 3 rows

Governance is coded separately from use: a report can mention AI policy, oversight, or risk-register work without showing that the organisation uses AI itself.

Governance Reports % AI-positive
formal 42 12.21
informal 28 8.14
none 274 79.65
Table · Chapter 5 Governance by Practice Maturity 18 rows

Own-use documents grouped by ordered maturity and whether governance is mentioned.

AI practice maturity Governance Reports Entities
Absent No governance mentioned 3 3
Absent Informal oversight 1 1
Absent Formal governance 1 1
Assessing opportunities No governance mentioned 27 23
Assessing opportunities Informal oversight 9 8
Assessing opportunities Formal governance 4 4
Governed intent No governance mentioned 13 7
Governed intent Informal oversight 1 1
Governed intent Formal governance 13 12
Limited trial No governance mentioned 11 11
Limited trial Informal oversight 1 1
Limited trial Formal governance 2 2
Routine use No governance mentioned 35 28
Routine use Informal oversight 5 4
Routine use Formal governance 9 8
Embedded practice No governance mentioned 1 1
Embedded practice Formal governance 4 4
Multiple stages Formal governance 3 3
Table · Chapter 5 AI Framing 4 rows

Opportunity and risk framing come from explicit opportunity/risk language in AI-positive reports.

AI framing Reports % AI-positive
opportunity_led 150 43.6
neutral 107 31.1
balanced 73 21.22
risk_led 14 4.07
Table · Chapter 5 Investment Intent 4 rows
Investment intent Reports % AI-positive
not_signalled 177 51.45
increase 108 31.4
maintain 41 11.92
uncertain 18 5.23
Table · Chapter 5 Primary Enablers 8 rows
Primary enabler Reports % AI-positive
none 183 53.2
capability 78 22.67
governance 37 10.76
partner 23 6.69
budget 10 2.91
demand 9 2.62
data 3 0.87
procurement 1 0.29
Table · Chapter 5 Primary Barriers 6 rows
Primary barrier Reports % AI-positive
none 334 97.09
governance 6 1.74
legacy_systems 1 0.29
culture 1 0.29
unclear 1 0.29
budget 1 0.29

Source snapshot: reports/nfp_ai_investigation/current/REPORT_CURRENT_DRAFT.md. Tables and charts are regenerated by reports/nfp_ai_investigation/build_current_report.py.