Steve Holyer, Data Platform Leader EMEA North, Informatica
Across a whole range of industries, agentic AI is ushering in a new era of autonomous intelligence and decision making.
AI agents are showing promise in transforming workflows that process significant amounts of real-time data. In healthcare, for example, they can reduce administration inefficiency, provide initial diagnoses, interpret test results, and forecast prognoses, freeing doctors to focus on treatment plans and patient care. In sectors like supply chains and pharmaceuticals, agentic AI can rapidly process vast streams of data, creating a whole new class of operational intelligence and efficiency.
But much of its potential remains untapped. Today, AI excels at structured, repetitive tasks but is not yet ready to manage complex tasks without human oversight. The question for businesses is, how can they close the gap between what AI can do and what it could do?
READ MORE: Allstate Celebrates Future Innovators Competition in Derry and Strabane
A double-edged sword
The answer lies in the data itself. Data is both the fuel for AI success and the bottleneck getting in the way of it. Simply put: clean, trusted and governed data is essential. Poor quality data will always lead to poor-quality results. Even the most advanced autonomous AI agent is powerless to overcome the limitations of the information that feeds it. And with AI regulation raising the bar for transparency, accountability and risk, getting it wrong could prove costly.
To unlock truly transformative AI projects, organisations must do more than simply access data – they need to actively govern it. That means ensuring data is standardised across systems, continuously refreshed, properly classified, and accessible in real time. Only then can AI agents operate with the accuracy, context, and agility required.
The real hero
The true catalyst behind agentic AI’s success is often overlooked: robust data management. It is the foundation on which the agentic AI revolution is built, the ground without which intelligent enterprises simply cannot scale.
Data leaders tell us that data quality, completeness, and readiness are their biggest obstacles when deploying AI initiatives. High-capacity, scalable, intelligent data management solves these challenges by delivering clean, trusted and accessible data to front-facing models at scale. This transforms raw data into a strategic asset and establishes data management not as a support function, but as a critical driver of AI value.
Building the structure of success
Organisations must first develop powerful capabilities in data integration, quality, and governance. They need to bring together enterprise-wide data into a single view, making it available for continuous checks and refinement. This ensures all data that’s passed forward is relevant, up-to-date, and in the right format, enriched with standardised metadata, and
sourced appropriately. Crucially, this results in data that is managed responsibly and ethically, a core expectation as AI systems increasingly influence decisions.
It may sound counterintuitive, but AI can help here. The same technology that relies on high-quality data can also be used to improve it. AI can resolve long-standing data challenges such as mismatched records and outdated entries. By using AI to cleanse and enrich datasets, organisations can escape the long-standing cycle of poor inputs undermining AI performance. This creates a virtuous loop in which smarter systems continuously refine the right data, data that is not only clean, but also fully understood, contextualised and governed.
Alongside improving data quality, it’s essential to maintain strict controls over how it is stored, shared, and accessed. Strong data governance is essential for regulatory compliance, especially as global AI oversight increases and ethical expectations rise. To move forward safely, organisations need a trust-by-design operating model. That means a holistic view of the data feeding AI agents, with clearly allocated ownership and a data catalogue run out of a centralised agentic AI data and controls hub.
READ MORE: UK backs next-generation satellite communications with £6.9 million investment
However, trust and governance aren’t only about how way data is held. These principles also depend on explainability and traceability – understanding not just what an AI agent has decided, but why. Organisations can no longer generate an outcome without being able to explain how it was derived. That requires embedding semantic controls and governance upstream, at the data management stage, so that context, provenance, lineage, and decision logic can be accessed throughout the process. This not only ensures compliance but also lays the foundation for competitive advantage.
Brought together, these capabilities enable organisations to reliably fuel their AI models with governed, well-understood data, empowering agentic AI systems to learn, adapt, and evolve with confidence.
Seizing the opportunity
By investing in scalable, high-capacity data management, organisations can overcome the challenges that hold back agentic AI deployment and harness its full potential. As autonomous systems become ever more capable, their effectiveness hinges on the quality of the data they process – making robust data management systems a critical cornerstone of any AI initiative. The future of AI will be bright, but only when its foundations are strong, and that every outcome can be understood, explained and trusted.
Sync NI's Autumn 2025 magazine explores how Northern Ireland is moving AI from pilot to production
This issue features exclusive insights from industry leaders on real-world AI implementation, responsible adoption, enterprise infrastructure, and how data-driven innovation is transforming businesses across the region.
Read the Big Data Special Edition free online →
Stay connected with NI's tech community:
Subscribe to our newsletter – Get the latest tech news, job opportunities, and events delivered to your inbox
Visit Sync NI – Your daily source for Northern Ireland technology news
