AI has moved from experimentation into the operating model of business. But widespread use should not be confused with enterprise transformation. The next phase will be decided by whether organisations can turn AI deployment into measurable productivity, better decisions and new products—while managing risk.
Adoption is no longer the question
Stanford HAI’s 2026 AI Index, drawing on McKinsey’s 2025 survey, reports that 88% of surveyed organisations used AI in at least one business function, up from 78% in 2024. Regular generative AI use reached 79%, compared with 71% a year earlier. These figures point to a rapid shift from isolated pilots to broader organisational use.
Yet the same research highlights a more nuanced picture: AI-agent deployment remained in the single digits across nearly all business functions. In other words, using AI to assist people is becoming common; delegating complex, multi-step work to autonomous systems is still an emerging capability.
The next gap is execution
For leadership teams, the strategic question is changing. It is no longer simply which model or assistant to buy. It is which workflows should be redesigned, what data those workflows can safely access, where human approval remains essential, and how the organisation will measure outcomes.
That requires investment in data quality, integration, employee training, security, evaluation and governance. Without these foundations, organisations risk accumulating disconnected AI tools that generate activity without changing business performance.
Three priorities for the next phase
- Move from pilots to priority workflows. Select high-value processes where AI can reduce cycle time, improve service or support better decisions.
- Build trust into deployment. Establish clear accountability, access controls, evaluation standards and human oversight.
- Measure outcomes, not usage. Track quality, cost, speed, customer impact and risk—not just the number of employees using AI.
The Metamorph perspective
As AI becomes embedded in products and business operations, communications will also need to mature. Companies must explain not only what their AI can do, but where it creates value, how it is governed and what makes its approach credible. In a crowded market, evidence-led communication can help distinguish real capability from inflated claims.
The organisations that lead the next AI cycle may not be those that adopt the most tools, but those that redesign work around them most effectively.
Sources
Explore related topics