The transition towards an Agentic Enterprise—where humans and AI agents collaborate—continues to develop within the UK. Organisations currently use an average of 13 AI agents, with this number expected to double within two years.
UK IT leaders report orchestration and governance challenges. At present, over half of AI agents operate in isolation rather than as part of a unified multi-agent system. This can result in disconnected workflows, duplicated automations, and potential risks related to shadow AI.
Research indicates that API-driven architectures are being used as a method to connect, orchestrate, and govern multi-agent systems across platforms.
As AI adoption increases, AI agents are moving from early-stage use cases to broader use in enterprise productivity. UK IT leaders are focusing on different types of agentic solutions and the development of communication protocols to manage them.
Diverse Development: Organisations use a mix of approaches to build AI agents, including prebuilt SaaS agents, embedded platform agents, and in-house development.
Protocol Adoption: Organisations are exploring and adopting a range of protocols to manage and connect AI agents.
While AI agent adoption is high, supporting infrastructure is often not fully integrated for multi-agent use cases that require coordinated access to enterprise data.
App and Agent Sprawl: Enterprises use an average of 796 applications, with around one-third integrated.
Integration Hurdles: Reported challenges include risk management, compliance requirements, lack of expertise, and integration of siloed systems.
Data-related barriers are also reported, including outdated infrastructure and disconnected systems. An estimated 22% of APIs are not governed, which is associated with concerns around shadow AI.
To address these issues, UK IT leaders are adopting unified approaches based on APIs. This approach is intended to support connectivity between AI tools and enable more coordinated multi-agent systems.
Connectivity Mandate: AI agent effectiveness is linked to data integration across systems.
Architecture Shift: There is movement towards more API-driven IT architectures to support integration between applications, data, and AI.
