Deloitte just published its annual State of AI in the Enterprise report, and one number stands apart from everything else in it: only 20% of surveyed organizations say their talent is highly prepared for broad AI adoption. That sits next to another number, which is that around 60% of workers now have access to sanctioned AI tools. One year ago, fewer than 40% did. Companies expanded tool access by roughly 50% in twelve months. Readiness did not follow. The split between tool access and workforce readiness is the number worth measuring, and it is the same problem that determines whether your AI investment pays off or disappears into a line item.
What did Deloitte actually measure?
The State of AI in the Enterprise: The Untapped Edge is Deloitte's annual survey of business and IT leaders directly involved in their organizations' AI work. The 2026 edition surveyed 3,235 leaders across 24 countries and six industries, including financial services, manufacturing, life sciences, and technology. Respondents ranged from board and C-suite members down to vice presidents and directors. The survey was conducted between August and September 2025, with findings published in early 2026.
The report's title captures the thesis directly: organizations are standing at the untapped edge of AI's potential. They have the tools. A growing share have the strategy. What they are missing is the workforce capability to turn both into results. Deloitte named it "From Ambition to Activation," which describes the distance between knowing AI is important and actually building an organization that can use it well.
The numbers are specific and worth sitting with. Workforce access to AI tools grew from fewer than 40% to around 60% of workers in a single year. That 50% expansion in tool access was deliberate. Companies made a conscious push to get more employees inside AI-enabled workflows. But among those workers who now have access to sanctioned AI tools, fewer than 60% use those tools in their daily workflow. That figure has not changed from the prior year. The tools arrived. The behavior did not change at the same rate.
of surveyed organizations say their talent is highly prepared for broad AI adoption, according to Deloitte's survey of 3,235 enterprise leaders across 24 countries.
Source: Deloitte, State of AI in the Enterprise 2026Why access and ability are not the same thing
Tool access and talent readiness are two different metrics. The first is a procurement decision. The second is a capability question. Buying a license and distributing software does not train anyone to use it well, let alone to integrate it into the actual structure of how the work gets done.
Deloitte put the question directly to leaders: what is the biggest barrier to integrating AI into existing workflows? The top answer was insufficient worker skills. Not data quality. Not infrastructure. Not governance. Worker skills. That answer comes from the same leaders who, in the prior question, said they expanded tool access by 50% in one year. The irony is built in: the primary response to the AI opportunity was expanding access, and the primary barrier to results is still capability.
The worker sentiment data adds the other dimension. Among non-technical workers (the people who make up most of any company's workforce), 13% are highly enthusiastic about AI and 55% are open to exploring it. That is 68% of your team that is at least willing. But 21% prefer not to use AI unless required, and 4% actively distrust it. The 55% who are open but uncommitted are the critical group. They have not said no. They also have not been given a clear path. Without a structured way to build skill and measure progress, that open-to-exploring posture stays open-to-exploring indefinitely.
What do companies actually do when they respond to this?
Deloitte asked leaders how their organizations are adjusting their AI talent strategy. The most common responses were: educating the broader workforce to raise overall AI fluency (53%), and designing and implementing upskilling and reskilling strategies (48%). Farther down the list: redesigning career paths (33%), assessing changes to skill supply and demand (30%), and combining or restructuring organizations based on new AI usage patterns (30%).
The priority order is telling. Companies are reaching for training first, restructuring last. Education is the easiest lever. You can point to a course completion rate and call it progress. Redesigning how work actually flows, which is what eventually determines whether AI produces results, is harder and slower. The data Deloitte collected reflects that: 66% of organizations are achieving productivity and efficiency gains from AI adoption. But only 25% of leaders say AI is having a transformative effect on their companies, and only 34% of companies say they are using AI to deeply transform the business. Broad gains, narrow transformation. A major part of the distance between productivity gains and business transformation appears to be whether organizations are redesigning work around AI, or simply giving people better tools for old workflows.
Tool access is a procurement decision. Talent readiness is a capability question. Buying licenses does not train anyone to use them well.
What the 20% figure actually means for mid-market companies
Deloitte's 3,235 respondents skew toward larger enterprises: global companies with complex AI programs and dedicated technology leadership. Mid-market companies in Indiana and across the Midwest are generally at an earlier stage than the survey average. That means two things.
First, for many mid-market companies, that 20% "highly prepared" figure may be better understood as a warning sign, not a comfortable benchmark. If the largest companies in 24 countries, many with dedicated AI officers and multi-year transformation programs, are only at 20% talent readiness, mid-market companies that launched AI tool access in the past twelve months are generally earlier in the curve.
Second, the actions companies are taking (focusing on education rather than redesigning how work flows) are the actions that will continue to produce that 66% productivity result and not the 25% transformation result. The distinction is worth understanding if you are paying for AI investment out of operating budget and expecting it to show up in revenue or cost structure.
The Deloitte data also breaks down organizational readiness across four dimensions: governance (30%), technical infrastructure (43%), data management (40%), and talent (20%). Talent readiness is the lowest of the four, by a significant margin. Companies have put more effort into building the technical and data foundation than into building the human one. That sequence makes sense if AI is a technology project. It stops making sense the moment AI becomes a workforce project, which is where most companies now sit.
of organizations report productivity and efficiency gains from AI adoption. But only 25% say AI is having a transformative effect on their company.
Source: Deloitte, State of AI in the Enterprise 2026Where The 7 Levels of AI Proficiency fits into this
The Deloitte finding is a diagnosis without a prescription. The report tells you that 80% of organizations are not where they need to be on talent readiness, and that the dominant response (general training) is not moving the readiness number fast enough. It does not tell you where any specific person or team sits, or what they need to do next.
That is the problem The 7 Levels of AI Proficiency was designed to address. The framework places individuals and teams on a seven-stage scale of AI capability, from Level 1 (AI Aware, where someone understands AI exists and has basic awareness of its applications) through Level 7 (AI Orchestrator, where someone is designing multi-agent systems and managing AI-driven workflows at scale). Each level has specific, measurable behaviors attached to it, and each level points to a concrete next step.
The 20% figure in the Deloitte report, described in population terms, likely maps to the upper levels of that scale. People operating at Level 4 or above in The 7 Levels of AI Proficiency are actively building workflows, integrating AI into recurring processes, and producing measurable output changes. The 55% who are "open to exploring" AI are probably clustered around Levels 2 and 3: they have tried AI tools, they see the value, but they have not yet built consistent habits or workflow integrations. The 21% who prefer not to use AI unless required are likely at Level 1 or below.
What the framework adds beyond a vocabulary is a shared language for what "ready" actually means, and a structured path between where someone is and where the business needs them to be. Without that, "we are running training programs" becomes the answer to every question about readiness, because there is no other way to measure it.
The free assessment at assess.launchready.ai takes about 10 minutes and places each person across the seven levels. For companies working through the Deloitte problem (lots of access, uncertain readiness), the assessment is the measurement instrument that turns "we think we're okay" into "here is where each person actually sits."
What the agentic AI data adds to this picture
One more dimension from the Deloitte report is worth naming, because it changes the timeline pressure on the readiness question. Deloitte found that 85% of companies expect to customize agents to fit the unique needs of their business. Separately, 74% of companies expect to use agentic AI at least moderately within two years. Agentic AI (systems that can plan, act, and complete multi-step tasks with some degree of autonomy) is expected to have the highest impact in customer support, with supply chain management, R&D, knowledge management, and cybersecurity also seen as high-potential areas.
But only 21% of those companies report having a mature governance model for agents. The readiness challenge extends well beyond teaching someone to use a chatbot. Building capability to work alongside, manage, and verify AI systems that are making decisions and taking actions inside business workflows requires a different skill set from what most AI training programs currently build.
In The 7 Levels of AI Proficiency, this is the difference between Level 3 (Critical Thinker, where someone uses AI tools effectively in their own work) and Level 5 and above (where someone is managing AI outputs, evaluating agent behavior, and designing systems with oversight built in). The current training focus on raising overall AI fluency builds Level 2 and 3 skills. The agentic AI challenge requires Level 4 and 5 thinking, and most companies have not started building that yet.
of companies expect to customize agents to fit the unique needs of their business. Only 21% have a mature governance model for those agents.
Source: Deloitte, State of AI in the Enterprise 2026Three things to do with this data this week
The Deloitte report is a credible, large-sample signal about where enterprise AI actually stands. Here is how to use it practically.
Measure before you train.
The most common response in the Deloitte data was "educate the workforce." But education without a baseline is guesswork. Before adding another training program, measure where your team is. A structured assessment, run across your leadership team and their direct reports, tells you where the capability is concentrated and where it is absent. Training without measurement is activity. Training with measurement is progress.
Separate access from readiness in your reporting.
If your AI update to the leadership team counts software licenses and course completions as the primary metrics, you are measuring access and activity, not readiness. The Deloitte data is useful precisely because it disaggregates the two. Build that distinction into your own reporting: how many people have access, and how many people are actually using AI in ways that change how the work gets done.
Identify the 55% and build for them specifically.
More than half of your team is open to AI but not yet committed. These are not resisters. They are people waiting for a clear path. General training does not give them that. A structured proficiency framework, with a clear next step for someone at each level, does. The people who move from "open to exploring" to "using it every day" are the ones who will move the outcome from 66% reporting productivity gains to 25% calling it transformative.
Related reading: Level 4: Commander in the 7 Levels of AI Proficiency.
Sources
- Deloitte. "From Ambition to Activation: Organizations Stand at the Untapped Edge of AI's Potential." State of AI in the Enterprise 2026 Press Release.
- Deloitte. "The State of AI in the Enterprise: The Untapped Edge." 2026 AI Report (full report page).
- BigDATAwire. "Deloitte's State of AI 2026: Why Enterprise Execution Is Falling Behind Adoption." March 3, 2026.
- Painter, Harrison. "What Is an AI Capability Audit?" LaunchReady.ai Insights.
- Painter, Harrison. "AI Proficiency vs AI Literacy vs AI Fluency: The Difference Is Load-Bearing." LaunchReady.ai Insights.
- Painter, Harrison. "Half of Your Team Uses AI. Only 12% Say It Changed the Work." LaunchReady.ai Insights.
Frequently Asked Questions
What did Deloitte find about AI talent readiness in 2026?
Deloitte's State of AI in the Enterprise 2026 report, based on a survey of 3,235 business and IT leaders across 24 countries, found that only 20% of surveyed organizations say their talent is highly prepared for broad AI adoption. This sits against a backdrop where around 60% of workers now have access to sanctioned AI tools, up from fewer than 40% the year before. The split between tool access and actual workforce readiness is the central finding.
Why do companies have strong AI tool access but weak AI readiness?
Companies expanded AI tool access by roughly 50% in a single year, but talent readiness did not move at the same rate. The primary reason, according to Deloitte's 2026 survey, is that insufficient worker skills are the biggest barrier to integrating AI into existing workflows. Most organizations responded by adding training programs rather than changing how work is structured. Education alone, without role redesign or proficiency measurement, does not move the readiness number.
What is the biggest barrier to AI integration at work?
According to the leaders in Deloitte's 2026 State of AI survey, insufficient worker skills are the top barrier to integrating AI into existing workflows. This outranked technical infrastructure challenges, data management issues, and governance concerns. The finding points to a workforce capability problem, not a technology problem.
What percentage of workers actually use AI tools in their daily workflow?
Among workers who have been given access to sanctioned AI tools, fewer than 60% use those tools in their daily workflow. This figure has remained largely unchanged from the prior year despite a major push to expand access. Having the tool available and using it consistently are two different outcomes.
What is The 7 Levels of AI Proficiency and how does it address the readiness shortfall?
The 7 Levels of AI Proficiency is a measurement framework developed by Harrison Painter at LaunchReady.ai that places individuals and teams on a seven-stage scale of AI capability, from Level 1 (AI Aware) through Level 7 (AI Orchestrator). The framework gives companies a concrete way to measure where their workforce actually sits, not just whether employees have access to tools. The free assessment at assess.launchready.ai takes about 10 minutes and produces a placement across the seven levels.
What did Deloitte find about AI productivity gains in 2026?
Two-thirds of organizations (66%) in Deloitte's 2026 survey reported productivity and efficiency gains from AI adoption. However, only 25% of leaders said AI is having a transformative effect on their company, and only 34% of companies say they are using AI to deeply transform the business. Broad adoption and transformative results remain separated by execution capability.
How do worker attitudes toward AI vary in 2026?
Deloitte's 2026 survey found that among non-technical workers, 13% are highly enthusiastic about AI, 55% are open to exploring it, 21% prefer not to use it unless required, and 4% actively distrust it. The largest group, more than half, is open but not yet committed. These workers are the primary audience for readiness programs: they have the motivation to learn but have not yet been given a clear path.
What should a CEO do after seeing the Deloitte AI readiness data?
The first step is measurement. Buying AI tools and running general training sessions does not tell you where your team actually sits on the capability scale. A structured proficiency assessment, like The 7 Levels of AI Proficiency framework, places each person on a concrete scale and shows you where the readiness shortfall is concentrated, by role, by team, by skill category. From there, targeted training (not blanket courses) closes the specific deficits that matter for your business.
Find your AI Proficiency level
The free 7 Levels assessment places you across seven stages of AI capability. Under ten minutes. Research-backed scoring.