Irish businesses are keen to expand their agent AI, with 70% planning to increase their AI budget by 2026, according to our AI Agent Survey. But adoption has been slow, with only 9% reporting widespread use of AI agents, compared to 52% in the US, and 83% still in the research or testing phase. This gap highlights the need for clear evidence and standards to support investments.
There was a lot of interest in agent deployments last year, but many did not quite meet expectations in terms of realized value. In some instances, implementations focused on experimentation rather than high-impact applications, which limited tangible results. When organizations were asked to demonstrate that their agents were actively delivering measurable benefits, it was often difficult to provide clear examples, highlighting the gap between early adopters and proven results.
We expect that to change in 2026. We now know what good agent AI looks like. It has proof points like benchmarks that track value that is important to the business, whether it's related to financials (profit and loss impact), operations (market differentiation), or employees and trust. Instead of siled efforts, we have a centralized platform for deployment and monitoring that leverages a shared library of agents, templates, and tools. Before each deployment, the agent is tested, defects are fixed, and a working demo is created for future users to try. This allows agents to provide feedback and start trusting your agent's capabilities.
Agents are deployed as part of an entirely new workflow, with clear steps for human initiative, review, and monitoring, and people with training and incentives to collaborate with agents and provide oversight.
Built-in monitoring also includes different agents that check each other's work, and in high-risk scenarios these agents can come from different model providers.
Because agents can automatically document their decisions and actions, continuous monitoring is highly effective for tracking adoption and performance, quickly remediating errors, and building stakeholder trust.
Today's agents are imperfect, but new technologies generally are. 2026 could be the year for agents to shine, as companies know how to focus and centralize deployments based on real-world benchmarks.
