Over the past few years, organisations have had to get comfortable with experimenting. The implementation of enterprise AI has seen most businesses invest in both the technology and the upskilling needed to bring it into their workflows. The challenge now is scaling adoption across the organisation.
Closing the gap between AI ambition and execution requires enterprises to reimagine work, help people build the fluency to work effectively with AI, and adapt to a fundamentally faster pace of innovation, while translating adoption into measurable business value.
Peter Drucker warned that "the large organisation has to learn to innovate, or it won't survive". With AI, learning to innovate increasingly means learning to adopt.
Despite heavy investment in generative AI platforms, scaling these initiatives to generate enterprise-wide returns remains a formidable hurdle.
While organisations are increasingly recognising the potential of AI, translating that potential into meaningful business value remains a challenge. The distinction lies in the ability to move beyond isolated initiatives and embed AI across the enterprise. The execution gap is no longer about proving the technology works; it is about scaling it.
Leaders must recognise that deploying more tools does not equate to adoption. Success requires identifying exactly where AI should change work and managing an organisation whose teams must now adapt to new ways of working.
Once businesses identify how they want AI to change how they work, they must empower their workforce to deploy it successfully.
However, our research shows that only 28% of respondents report having an enterprise-wide skills framework or an evolving skills taxonomy for an AI-enabled workforce.
Providing access to platforms and high-level training isn’t enough to break the pull of legacy workflows. These steps won’t tell workers which parts of their role an agent can support, or what tasks still require human judgement.
To achieve true enterprise value, AI must be embedded directly into roles, processes, and value chains. This means moving beyond broad training programmes to highly targeted, hands-on adoption that addresses the realities of daily work.
Focused, role-specific mentoring is essential. When TCS challenged its internal functions to adopt AI, an initial broad immersion session generated interest but did not shift established behaviours. Making the technology stick meant narrowing the focus, spending four to six months mentoring a single, specific role.
This helped employees identify exactly where an AI agent could assist them using their own daily data – such as specific spreadsheets and files.
Adoption at scale requires this level of hands-on mentoring. When employees are supported to ‘bend’ AI agents to solve their own routine problems, they build genuine fluency and unfreeze legacy mindsets.
As skills and familiarity with the technology improve, organisations must examine how they achieve true transformation. If AI is only applied to existing processes and to accelerate individual tasks, it will yield minimal returns.
Greater value comes from building on the fluency developed through mentoring and rethinking how entire processes should work, redesigning end-to-end workflows to capture true operational efficiency.
In recent conversations with leaders from a wide range of industries, we explored how businesses can advance beyond updating their infrastructure to deploy AI at an intelligence level. Doing so requires a systemic approach, tailored to specific organisational requirements.
When done well, there is great opportunity to unleash true competitive advantage. The greatest value comes not from applying AI to existing ways of working, but from fundamentally redesigning workflows to unlock new levels of efficiency, productivity and innovation.
We have observed that as teams become more fluent in AI, the pace of work is compressed, and enterprises now have three distinct operating lanes to consider.
For leaders, the challenge is orchestrating all three speeds in harmony. A team responsible for a critical core system cannot set aside necessary controls just because another team can produce a solution quickly. They must create frameworks that allow them to assign work to the right lane based on a strict assessment of stability versus speed requirements.
Interestingly, AI itself is becoming a critical tool in navigating this complexity. Our recent research shows that decision-making is the area that executives say is experiencing the greatest AI-related change – meaning leaders can increasingly rely on AI to generate insights, predictions, and recommendations.
Ultimately, accelerating innovation is not the same as insisting everything happen faster. Effective governance is a strategic leadership choice: knowing exactly where speed is appropriate to capture a market opportunity, and where deliberate control and API gateways must be maintained to protect the enterprise.
The promise of scaled AI lies in its ability to radically optimise established operations – reducing effort on routine work, accelerating delivery, and improving customer experiences.
Last year, in collaboration with the Peter Drucker Society Europe, we explored the qualities leaders need to empower their talent and support change. By decentralising decision-making, businesses can tap into colleagues’ experience and build an organisation where new mindsets are embedded, top to bottom.
This means connecting the experience of the people doing the work to decisions about whole processes. Doing so requires a willingness to optimise legacy core operations, which, in turn, generates the cost savings needed to fund broader enterprise modernisation.
As the 2026 Drucker Forum theme explores, next-generation innovation depends on human ingenuity. AI gives enterprises the capacity to work differently, but turning that capacity into value is a distinct leadership task.
Outcomes must be assessed, not assumed. Success should not be measured by the number of platforms deployed or agents created, but by strict operational metrics: reduced process cycle times, improved output quality, and realised financial impact. The real question is how innovation at different speeds can best amplify human ingenuity to serve the business as a whole.