Unlocking the benefits of this technology hinges on eliminating data siloes and unifying information across an organisation
Interest in autonomous AI tools is accelerating, as businesses look to streamline operations and enhance customer experiences. Among the most promising of these innovations is agentic AI — systems that can independently perform tasks and make decisions with minimal human input. But while adoption is rising fast, many organisations are overlooking a critical foundation: high-quality, unified data. The true impact of agentic AI depends not just on the technology itself, but on a company’s ability to eliminate siloes and structure information in a way the AI can understand and trust.
According to KPMG’s ‘AI Quarterly Pulse Survey’, 65% of organisations surveyed in Q1 2025 said they were piloting AI agents, increasing from 37% the previous quarter.
According to Yohan Lobo, Senior Industry Solutions Manager at M-Files, before businesses rush to integrate agentic AI, they must first ensure they have an ordered, trustworthy bank of internal data for these tools to leverage.
Yohan said: “Exploration into agentic AI has gathered significant pace in recent months, with companies eager to benefit from the potential improvements in productivity and customer satisfaction this innovation is capable of delivering.
“AI agents are proving so popular because of the efficiency gains that can be achieved in both the front and back office. They can support employees with everyday tasks, picking up the burden of manual work and creating more time for staff to spend on activity that adds value.
“Additionally, agentic AI can be employed to streamline customer experience, responding to queries, reducing response times and offering user-specific personalisation. Organisations that implement this technology successfully on both these fronts will see considerable improvements in efficiency and quick return on investment.
“However, the integration of AI agents isn’t as simple as developing this technology or partnering with a vendor, then skipping straight to the deployment phase. If the data underpinning the AI tool isn’t ordered and cohesive, the agent will be unable to deliver trustworthy responses.
“Therefore, the first step for businesses considering how they can implement agentic AI is to conduct an audit of their data, identifying where the information at their disposal is fragmented and where they’re operating across disparate systems. Once this data has been carefully curated, organisations can begin to benefit from the potential of AI agents.
“People are referring to data as the modern equivalent of oil, but this digital fuel first needs to be refined before it can power new tools like agentic AI. Once the platform is set, the use cases are extensive. Take the insurance industry as an example. Here AI agents can handle a vast range of activity, from customer servicing to claims processing.”
Yohan concluded: “Agentic AI represents a host of possibilities capable of transforming business performance. The only way to access these improvements is to focus on data to ensure that these solutions are accurate and reliable. If not, employees and customers alike will neither trust, nor utilise AI agents.”
