Imagine you are working in an office where some colleagues refuse to touch artificial intelligence, others experiment with it casually, and a few design entire systems around it. This variation in behaviour reflects what consulting firm KPMG describes as different levels of AI competence within a workforce. The framework helps organisations understand how individuals interact with AI tools, ranging from complete disengagement to full technical development. Recognising these levels matters because artificial intelligence is reshaping professional workflows by assisting with research, writing, analysis, and automation. Understanding where employees sit on this spectrum allows organisations to identify training needs, reduce resistance to new technologies, and unlock productivity gains as AI adoption expands across industries.
KPMG’s framework describes four broad levels of AI competence that reflect how individuals interact with the technology in their daily work:
• Level 0 — AI Disengagement: Individuals avoid or ignore AI tools entirely. They may be sceptical, fearful of the technology, or believe it is irrelevant to their role. Tasks that could be assisted by AI are performed manually.
• Level 1 — AI Engagement: Users experiment with AI tools casually but without integrating them into structured workflows. They may ask simple questions, generate drafts, or summarise documents using AI systems.
• Level 2 — Non-Technical AI Application: Individuals integrate AI into regular workflows even though they do not build AI systems themselves. Activities may include prompt design, automation using digital tools, and using AI for research, planning, writing, or decision support.
• Level 3 — Technical AI Development: Specialists design, build, or customise AI systems. This includes coding applications, training models, and building data pipelines or AI agents that others can use.
The first two levels illustrate early-stage interaction with AI technology. People at the disengagement stage typically maintain traditional manual workflows despite the availability of AI tools that could accelerate research or document preparation. In contrast, those in the engagement stage have begun experimenting with AI systems but still treat them primarily as simple assistants or search tools. Organisations often observe these behaviours during the early phases of AI adoption, when awareness of the technology grows faster than structured implementation within professional workflows.
The third and fourth levels represent deeper integration of artificial intelligence into organisational capability. Non-technical AI application involves embedding AI tools into routine processes such as content generation, workflow automation, and information analysis. For example, a professional might build an automated content pipeline in which AI generates written material, publishing systems distribute it, and analytics platforms monitor performance. Technical AI development goes further by creating the underlying systems themselves. Like driving a car without building the engine, many professionals operate AI tools effectively without constructing the underlying models.
Understanding these four levels provides organisations with a practical framework for workforce development. Studies from KPMG indicate that AI adoption across businesses varies widely, with employee confidence and training playing major roles in successful integration. By recognising where individuals fall on the competence spectrum, leaders can design targeted education programs and encourage responsible experimentation with AI tools. For readers encountering this framework today, a practical step is to assess current work habits and identify opportunities to move from occasional experimentation toward structured AI-assisted workflows, which research commonly associates with meaningful productivity improvements.