OutSystems survey finds AI governance gap in firms

OutSystems has published research showing a gap between business confidence in governing AI agents and their operational readiness to manage them. The survey found that only 26% of IT decision-makers say they are well prepared to resolve conflicts across AI agents, systems and automated workflows.

The research, conducted with CIO Dive and KPMG, surveyed 858 executives and points to growing pressure on companies as AI projects move from trials into wider use. While 82% of respondents said their organisations could keep pace with governing, integrating and monitoring AI agents, the findings suggest that confidence weakens when businesses face the practical demands of oversight and control.

European organisations stood out for their growing reliance on outside help. Some 40% said they expect to become more dependent on strategic partners and managed services as AI costs rise, compared with 31% in North America and 26% in Asia-Pacific.

The figures suggest spending concerns are shaping how companies approach AI adoption. In Europe, respondents also expressed stronger concerns about reliability, with 49% citing inaccurate, unreliable or unexpected agent behaviour as a major orchestration challenge, compared with 41% in North America and 35% in Asia-Pacific.

Governance gap

The study focused on agentic AI, in which software agents can act across applications and workflows with a degree of autonomy. It found that many companies still rely on established governance structures rather than creating dedicated controls for this type of system.

According to the survey, 83% of organisations use existing IT, security or compliance processes, specialised reviews, or centralised policies and monitoring to oversee AI agents. Only 5% said they had automated controls, audit trails and formal human review points in place.

The imbalance suggests many businesses are adapting older oversight methods to newer AI systems. It also points to a broader challenge for companies that want to deploy multiple agents across different tools, data sources and business processes without creating new operational risks.

Integration emerged as a central concern. Some 74% of respondents said integrating with existing applications and systems was a priority, 63% pointed to orchestration across systems, and 62% said it was important to keep the option of switching models, vendors and platforms over time.

Operational focus

The findings also suggest companies are measuring AI progress through internal operating results rather than customer-facing change. Error reduction was the most commonly cited measure of impact at 64%, followed by compliance and risk reduction, employee productivity, and cost reduction.

By contrast, only 12% of respondents said agentic AI had accelerated the launch of new products, services or customer experiences. That points to a more cautious phase of adoption, in which businesses are looking first for control, reliability and efficiency gains.

Even so, expectations for broader use remain high. Around 71% of executives said they expect applications to coordinate work across systems, teams and business processes, or support more complex multi-step workflows, within the next five years.

The report also highlighted concerns about the quality of enterprise data and context. More than a third of respondents, 35%, said data consistency and context sharing were major orchestration hurdles, suggesting many organisations still face basic information management problems as they expand AI use.

Control systems

OutSystems Chief Information Officer Tiago Azevedo said the headline figures show where the next phase of work will fall for many businesses.

"82% of executives say they can keep pace with governing AI agents, yet only 26% say they can resolve conflicts between them. That gap is where the real work happens. Having put agents to work across the business, I've learned that safety comes from building control into the platform, not reviewing it afterward. Enterprises that enforce policy while agents run will move faster, because they won't have to stop and ask if they can trust what they built," Azevedo said.

KPMG Managing Director Michael Harper said established software governance methods were not enough for dynamic AI systems.

"Assurance for agentic systems must be engineered from day one - you cannot simply inspect it after the fact. Traditional logging and procedural governance were built for static software, not dynamic AI. To bridge the gap between high confidence and true operational readiness, enterprises need the ability to execute policy-as-code and runtime guardrails. That gives businesses the autonomy to innovate while keeping end-to-end workflows secure, reliable and focused on business outcomes," Harper said.