Agentic AI Goes Mainstream In The Enterprise But 94 Per Cent Raise Concern About Sprawl

Nearly every enterprise is now using AI agents in some capacity, but governance is failing to keep pace, with the vast majority of organisations flagging AI sprawl as a growing source of complexity, technical debt and security risk.

That is the central tension in OutSystems’ 2026 State of AI Development report, which surveyed 1,900 IT leaders globally, including 527 respondents from the Asia-Pacific region spanning India, Australia and Japan.

From pilots to production

The report paints a picture of an enterprise landscape that has moved well beyond experimentation. Ninety-six per cent of organisations surveyed are already using AI agents in some form, and 97 per cent are exploring system-wide agentic AI strategies.

The shift from pilots to production is most visible in IT and software development, where time-to-value is easiest to measure. Thirty-one per cent of respondents indicated AI is already integral to their development practices, and another 42 per cent have embedded AI into specific phases of the software development lifecycle.

Generative AI-assisted development emerged as a leading method in markets including India and Australia.

As agents demonstrate value in development environments, 52 per cent of organisations now rely on a human-on-the-loop model, allowing systems to operate with reduced direct oversight while maintaining supervisory control.

Australia at the intermediate stage

Adoption maturity varies considerably by region. India stands out with some of the highest levels of advanced and expert agentic AI capability. Australia and Japan reflect a growing base of organisations at an intermediate stage of maturity, steadily moving initiatives from pilot to production.

Organisations in Brazil, Germany, the Netherlands, the UK and the US are also reporting intermediate progress. Financial services and technology organisations report the highest levels of production deployment globally.

Across the full survey, 49 per cent of respondents described their agentic AI capabilities as advanced or expert.

The governance gap

While adoption has accelerated, governance structures have not kept up. Ninety-four per cent of organisations reported concern that AI sprawl is increasing complexity, technical debt and security risk.

Only a small fraction of enterprises have established a centralised approach to agentic AI governance. Most are using agents across fragmented environments, with governance approaches that vary by team and region.

Architectural fragmentation is a related challenge. Thirty-eight per cent of organisations globally report mixing custom-built and pre-built agents, creating AI stacks that are difficult to standardise and secure. Just 12 per cent have implemented a centralised platform to manage sprawl.

Woodson Martin, CEO at OutSystems, framed the governance challenge as the next frontier for enterprises that have already solved for adoption.

“The transition from AI experimentation to measurable business outcomes is no longer a future state – it is our current reality,” Martin observed. “The findings in the State of AI Development Report reveal a fundamental shift where building software and building AI systems have become one and the same.”

“As organisations move toward a ‘system of agents’ model, the challenge is no longer just about adoption, but about creating a stable architectural foundation that can coordinate these complex intelligent systems to drive real-world productivity.”

What makes agentic AI different

Agentic AI represents a notable evolution from earlier applications of AI in the enterprise. Rather than simply generating content or providing recommendations, agentic systems are capable of autonomously executing workflows, making decisions and adapting in real time.

Gartner predicts that 40 per cent of enterprise applications will include task-specific AI agents by the end of 2026, an indication of how quickly autonomous systems are becoming embedded in enterprise software.

That speed of adoption is precisely what is creating the governance headache. When AI agents are deployed in isolated pockets across an organisation, each with different levels of oversight and different integration points, the result is a sprawling landscape that is difficult to monitor, secure and maintain.

Starting small, building muscle

For some organisations, the approach has been to start with tightly scoped projects before expanding.

Scott Finkle, VP of Technology at McConkey Auction Group, described the company’s strategy as deliberately incremental.

“Our approach to working with OutSystems for an agentic solution was to start with a small, well-defined project that we felt like we could get into production, and that would actually have an impact on the business,” Finkle explained.

“Our main goal of the project was to build some muscle for building AI projects moving forward.”

Standardisation as the next challenge

The report suggests that the enterprise AI conversation is shifting from “are we using AI?” to “how do we govern what we have built?”

With agents now operating across development environments, customer-facing applications and internal workflows, the lack of standardised architecture is becoming a practical problem. Security teams need visibility across agent deployments. Compliance teams need consistent governance. And IT leaders need confidence that the agents operating within their environments are doing so within defined boundaries.

OutSystems has introduced what it calls Agentic Systems Engineering, an open approach to AI development designed to help organisations build, manage and evolve governed agentic systems.

Whether enterprises adopt that specific framework or develop their own, the report makes clear that the window for ad hoc AI governance is closing. As agentic AI moves from experiment to infrastructure, the organisations that establish centralised control early are likely to avoid the technical debt and security exposure that come with sprawl.

Last Updated on April 22, 2026 by Nick Ross