Skip to main content
Welcome to Deloitte

If we have selected the wrong experience for you, please change it above.

Deloitte research: AI is being deployed more often, but does not deliver the same value everywhere

Key findings

  • AI is speeding up decisions and execution: 81% of surveyed organisations report progress in these areas.
  • Organisational change is a bottleneck: organisations' ability to redesign work and governance is hindering their ability to realise value from AI.
  • AI does more than make recommendation - it takes action: 40% of surveyed organisations allow AI systems to make decisions or take actions, to some extent, without human approval. In 23% of organisations, it is unclear who is responsible when AI leads to a negative outcome.
  • The returns from AI go beyond productivity: 28% of surveyed organisations report revenue growth as a result of AI, compared with 20% in Deloitte's January 2026 report. 

2 November 2026

The latest edition of ‘State of AI in the Enterprise’ shows how organisations are moving from experimenting with AI to applying it in their operations. ‘The State of AI in the Enterprise: Built Differently, How AI is Reshaping the Way Enterprises Work, Grow and Create Value’ is based on a survey of 3,500 executives, from directors to board members, across 25 countries and six industries. They are directly involved in their organisations' AI initiatives. The report describes AI's potential to drive change, how its use is developing and what leaders should consider as they continue to deploy AI. 

"Many organisations are getting a clearer picture of their total AI costs, but far fewer understand the underlying unit costs: where those costs arise and whether they generate enough value. This can erode returns. Leaders who can link the cost of a workflow to the value it generates, and combine this financial discipline with organisational redesign, may be better placed to gain a competitive advantage and scale AI sustainably." - Costi Perricos, Deloitte Global GenAI Business leader

The new AI bottleneck: working faster is not the same as working differently

AI is speeding up work: 81% of the organisations surveyed report faster decision-making and execution. 

At most of the organisation surveyed, AI is speeding up existing workflows without changing the work itself. The question is whether organisations can use that acceleration to work, make decisions and create value differently.

"AI makes processes faster, but working faster is not the same as transformation. Only 25% of the organisations surveyed say they are already making significant changes to their businesses. The real challenge is not to speed up yet another task, but to determine which processes need to be redesigned end to end. Only when people and technology work well together in this redesign can AI deliver real change and value." - Jorg Schalekamp, AI Lead Deloitte Nederland

Organisations can see their AI bill, but often not what is driving the costs

Greater visibility among leaders into total AI spending does not always make the underlying costs clear. Of the organisations surveyed, 71% track total spending on AI inference and APIs, 51% track token volumes and 44% track the cost per inference unit. 

This gap in visibility makes it harder to assess whether the value AI delivers outweighs its costs. As investment in AI increases and applications become more complex, leaders need insight into the costs per workflow and the usage driving those costs in order to manage spending and assess returns.

The returns paradox: benefits emerge at different times

The benefits of AI do not all become visible at the same speed; financial returns take longer. However, 28% of the organisation surveyed now report revenue growth driven by AI, up from 20% in Deloitte's January report. These organisations also see benefits beyond productivity and cost savings. This suggests that value comes not only from working faster, but also from working smarter. 

Job redesign is still evolving

Of the organisations surveyed, 60% have redesigned jobs around AI capabilities to some extent. However, only 15% report a significant shift in decision-making authority as a result of AI. 

Employees may have better information and make decisions more quickly, but that does not automatically give them the authority to act differently. Effective work redesign also requires changes to roles, responsibilities, approval processes, decision-making authority and the allocation of work between people and AI. 

From advising to acting: governance and safety must keep pace

As AI takes on a greater role in critical operational processes, safety is essential. Forty per cent of the organisations surveyed allow AI to make decisions or take actions, to some extent, without human approval. At the same time, 235 say it is unclear who is responsible when an AI system causes a negative outcome.

Governance and safety measures should match the degree of autonomy of an AI system and the potential consequences of its actions. Systems that take action may require more oversight than systems that only make recommendations. 

About the research

The State of AI in the Enterprise is now in its ninth year. The research builds on Deloitte's previous quarterly State of Generative AI in the Enterprise reports, which tracked the rapid adoption of generative AI in organisations. For this edition, 3,500 business and IT leaders were surveyed between May and June 2026 across 25 countries and six industries: consumer; energy, resources and industrials; financial services; life sciences and healthcare; technology, media and telecommunications; and government and public services.