Given the increasing role of artificial intelligence (AI) in enterprise over the last few years, recent research is reflecting a shift from experimentation to implementation. Organisations are now moving beyond isolated pilots to scale agentic AI – autonomous systems capable of coordinating work, routing decisions, and delivering business impact, according to research from Deloitte and KPMG.
KPMG International’s first quarterly Global AI Pulse survey shows that global enthusiasm for AI is unwavering, with leaders planning a weighted global average of US$186 million in AI over the next 12 months, and 74% saying that AI will remain a top investment priority even in the event of a recession. While most organisations (64%) cite AI as already delivering meaningful business outcomes, they face growing challenges – from measuring and quantifying value, to adapting governance models at the required speed, managing data privacy and cyber risks, and addressing workforce resistance. These risks and challenges are keeping many global organisations in the experimentation and piloting stage of AI implementation. On the bright side, a meaningful minority (11%) are gaining an edge through AI agent deployment, scaling across functions and beginning to coordinate them across work flows.
While the results of the Global AI Pulse survey indicate that leading organisations are moving beyond enablement to deploying AI agents to reimagine processes and reshaping decisions and work flows across the enterprise, the report reinforces that spending more on AI is not the same as creating value. Ultimately, there will be no agentic future without trust, and no trust without the governance that keeps pace with evolvement. It makes clear that sustained investment in people, training and change management is what allows organisations to scale AI responsibly and capture value.
The key findings of the KPMG survey are:
Separately, Deloitte has released Agentic AI – Lessons from the Real World, a thought leadership piece sharing how organisations can effectively move beyond experimentation and pilots to scale the use of AI. It builds on the 2026 State of AI in the Enterprise report, which highlights that organisations are now moving beyond pilots to scale AI meaningfully.
Here are the practical lessons to help leaders translate AI ambition into on-the-ground impact:
The core foundation for successful AI initiatives is trust, and trust should underpin all aspects of agentic AI design. It needs to be earned through clear human-agent decision boundaries, transparent reasoning, human feedback loops, audit logs and the delivery of consistently accurate results.
Agentic AI amplifies the process it enters; it does not fix a broken one. Successful agentic AI initiatives start with carefully identifying the processes where measurable returns can be realised, followed by work flow redesign to optimise the way people and agents work together to deliver improved outcomes. When value definition is clear and processes are well designed, agentic AI becomes tangible accelerators of business performance.
Autonomy does not replace accountability, and the use of AI agents elevates the importance of human expertise and judgement. Leaders need to guard against complacency and over-reliance, and protect long-term capability development to ensure that automation does not erode professional judgement and craft.
Success at an enterprise level requires a whole-of-organisation mindset from the start, disciplined focus on top-down business objectives, robust enterprise architecture, organisational governance, and building systems for scale – not show.
A key objective should be to unlock a tipping point of engagement and momentum at the enterprise level. This requires strong executive sponsorship and bringing the workforce incrementally on the journey through clear examples of AI removing friction, improving work processes, and delivering tangible value in real workplace settings.
Related articles on CA Lab:
