Deloitte: 81% of companies run faster with AI, only 25% have changed the business
Eighty-one percent of companies say AI has made decisions and execution faster; only 25% use it to deeply change the business through new products, reinvented core processes or new business models. That gap is the core finding of Deloitte’s State of AI in the Enterprise, a series now in its ninth year. The survey ran in May and June 2026 and covered about 3,500 director-to-C-suite leaders across 25 countries and six industries.
The returns are arriving, unevenly. Some 28% report revenue gains from AI, up from 20% in Deloitte’s January report. Sixty percent have redesigned jobs around AI to at least a moderate degree, but only 15% report a significant shift in decision-making authority. Faster work inside an unchanged org chart is still the dominant pattern — the constraint has moved from adoption to redesign.
Governance and cost measurement trail adoption. Forty percent let AI execute decisions or actions without human approval, at least moderately, while 23% say accountability is unclear when an AI system produces a negative outcome. On spend, 71% track total inference and API costs, but only 51% track token volume and 44% track per-unit inference cost. Per-unit cost is the number you need to compare one workflow against another.
Forrester’s Tracy Woo makes a related argument in Tokenomics: The Economics Of AI Value Realization: tokens are only one part of total AI cost of ownership, and infrastructure, data, software, labor, governance and operations can be larger cost drivers. Her advice is to measure value through productivity, revenue growth, customer adoption or process efficiency rather than spend alone.
For anyone shipping AI inside a company, the next step is concrete: pick the workflow, redesign who decides, name who is accountable when the system is wrong, and track unit cost against the outcome it produced. Adoption alone is no longer the outcomes metric.