Governance lags agentic AI adoption in the UK, says Salesforce

An orchestration and governance gap is emerging as enterprises deploy AI agents, according to research from Salesforce. In its Connectivity benchmark report, Salesforce-owned Mulesoft found that the organisations surveyed have 12 agents on average, with the number projected to climb by 67% over the next two years.

This is the vendor’s 11th iteration of the report, produced in collaboration with Vanson Bourne and with input from Salesforce’s Mulesoft partner Deloitte Digital. It surveyed 1,050 IT professionals from the US, UK and Ireland, France, Germany, the Netherlands, Australia, Singapore, Hong Kong and Japan between October and November 2025. The UK cohort comprised 100 respondents.

It found that 89% of the UK and Ireland organisations already deploy AI agents, but that half of the agents are compartmentalised into “silos”, so are not meaningfully deployed at an enterprise level. Only 54% of organisations have a centralised governance framework with formal oversight for their agents. Some 94% of the respondents agreed that AI agents will require new IT architectures to become more API-driven to succeed.

Enterprises are using 796 applications on average, with 33% of them integrated with each other, while 75% of the respondents are concerned that agents will introduce more complexity than business value.

The majority (97%) of the organisations said they have barriers to using data for AI use cases, with 35% identifying outdated IT architecture and infrastructure due to data silos and disconnected systems as being the top blocker. 

In the foreword to the report, Beena Ammanath, Global Deloitte AI Institute leader, said: “AI adoption speed has surpassed predictions from even a year ago, with 84% of enterprise CIOs believing that AI will be as important to their businesses as the internet.

“Now, on the cusp of the next era of AI – the agentic enterprise – 40% of organisations report having deployed autonomous agents to augment current processes and teams, with another 41% planning to deploy agents within the next year.”

In a press briefing on moving from single agent deployments to multi-agent systems, at which the research behind the Connectivity benchmark report was presented, Kurt Anderson, managing director and API transformation leader with Deloitte Consulting, expressed caution about the danger of shadow AI.

“There is an important leadership principle of putting knowledge in the hands of the people who can deploy it, who have the best information to make a decision and build an agent,” he said. “What’s different about this [agentic AI] revolution is we are embracing that knowledge and tech is lowering the bar to building your own capabilities…We can put guard rails and transparency and security in place and so get knowledge into the agents.

“That part is really cool compared with a past approach that required a more top-down, solid-hand type of governance. But we do have to learn from the past, and make sure we have the right governance. We know what happens when you let individual contributors loose in terms of building systems that lack security or reliability or run up a big licensing bill.”

In a post pointing to the report, Andrew Comstock, senior vice-president and general manager of MuleSoft, said: “The true success of an agentic enterprise isn’t found in the sheer number of agents deployed, but the overall effectiveness of those agents. We need to think about how they are discovered, governed and orchestrated to work together.

“As we move into this multi-agent era, the role of IT is evolving from managing silos to building a unified foundation as the central control plane that allows multi-agent systems to be safe, reliable and scalable.”