Data report
State of AI Use Cases
What 1377 documented AI deployments across 257 use cases say about where AI runs, how far it acts on its own, and how often organizations prove the result.
Computed live from the library on 3 October 2026. By Len Debets.
Finding 1: most AI results are claimed, few are proven
60% (n=1377) of public deployments disclose a measurable outcome. Of the 872 reported metrics, 56% come from vendors and 44% from the deploying organization itself.
- Vendor490
- Organization380
- Independent2
- Grade C788
- Grade B583
- Grade A4
- Grade D2
Finding 2: AI is moving from answering to acting
63% (n=1377) of public deployments belong to use cases where the AI completes work itself (a supervised or autonomous agent, or an agentic workflow), rather than only assisting a person.
- Supervised agent103
- Copilot87
- Assist51
- Autonomous16
- Production838
- Scaled298
- Pilot146
- Announced83
- Paused or rolled back12
Finding 3: where the evidence is
Public deployments by industry of the deploying organization, with the share that disclosed an outcome, and by region.
- Government and public sector227 (30% with outcome)
- Banking209 (59% with outcome)
- Insurance98 (53% with outcome)
- Retail and ecommerce95 (66% with outcome)
- Technology and software94 (89% with outcome)
- Telecommunications93 (67% with outcome)
- Healthcare80 (80% with outcome)
- Travel and hospitality62 (69% with outcome)
- Real estate47 (72% with outcome)
- Manufacturing46 (80% with outcome)
- Wealth and asset management37 (65% with outcome)
- Payments and cards36 (67% with outcome)
- Pharma and life sciences36 (47% with outcome)
- Media and entertainment35 (51% with outcome)
- Education35 (77% with outcome)
- Professional services35 (77% with outcome)
- Energy and utilities33 (64% with outcome)
- Automotive31 (48% with outcome)
- Logistics and transportation27 (44% with outcome)
- Capital markets15 (73% with outcome)
- Cross industry6
- North America681
- Europe349
- Asia Pacific163
- Global89
- Middle East38
- Latin America36
- Africa21
Finding 4: the building blocks
The AI patterns use cases are built from, and how the EU AI Act classifies them.
- Agentic workflow109
- Classification and routing90
- Prediction and scoring90
- Conversational agent76
- RAG knowledge assistant76
- Document processing71
- Summarization67
- Content generation63
- Anomaly detection56
- Voice agent35
- Recommendation and personalization27
- Computer vision27
- Translation17
- Speech analytics13
- Code generation7
- Synthetic data generation3
- Digital human1
- Depends on design123
- Minimal risk60
- Limited risk (transparency)56
- High risk18
The benchmarks with the most evidence
KPIs with at least three pooled public values across the whole library ("up to" values are listed on each KPI page, not pooled).
| KPI | Median | Reported range | Data points | Claimed by |
|---|---|---|---|---|
| Cycle time reduction | 55.5% | 7% to 97% Not pooled: up to 70%, up to 70%, up to 30%, up to 20% | 50plus 4 up to | 21 organization, 29 vendor |
| Productivity gain | 50% | 8.7% to 113% Not pooled: up to 90%, up to 87.5%, up to 80%, up to 30%, up to 16% | 40plus 5 up to | 15 organization, 25 vendor |
| Handling time reduction | 61.5% | 8.2% to 97% Not pooled: up to 85%, up to 67%, up to 50%, up to 20%, up to 5% | 36plus 5 up to | 11 organization, 25 vendor |
| Automation rate | 80% | 25% to 99% Not pooled: up to 70%, up to 50% | 37plus 2 up to | 13 organization, 24 vendor |
| Accuracy | 91% | 42% to 100% Not pooled: up to 99%, up to 90% | 33plus 2 up to | 18 organization, 15 vendor |
| Containment rate | 61.5% | 20% to 99.6% Not pooled: up to 88.5% | 28plus 1 up to | 12 organization, 16 vendor |
| Cost reduction | 45% | 2% to 95% Not pooled: up to 70%, up to 70% | 20plus 2 up to | 7 organization, 13 vendor |
| Employee adoption | 82% | 30% to 100% | 17 | 5 organization, 12 vendor |
About this report
Every figure is computed from the public records in the library at the date above, using the rules on the methodology page. The underlying data is available as open data under CC BY 4.0. The sample reflects what organizations choose to publish, which skews toward large organizations and successful projects.