EY helps clients create long-term value for all stakeholders. Enabled by data and technology, our services and solutions provide trust through assurance and help clients transform, grow and operate.
At EY, our purpose is building a better working world. The insights and services we provide help to create long-term value for clients, people and society, and to build trust in the capital markets.
In this episode of EY.ai Unplugged podcast series, we explore how organizations can prepare for the next evolution of Agentic AI systems powered by trusted, high‑quality data. Alexy Thomas, Partner, Technology Consulting, EY India, discusses why data is the single biggest factor in enabling intelligent, autonomous decision‑making and how poor-quality impacts outcomes, especially in case if Agentic AI. He also describes that high quality data drives smarter agents and in turn, AI continuously improves data quality – both complementing each other. He also shares characteristics of AI‑ready data and practical steps organizations can take to build a strong, scalable foundation for autonomous systems.
Key takeaways:
Trusted, high‑quality data is foundational for Agentic AI, directly shaping accuracy, decisions, autonomy and overall intelligence of AI systems.
Agentic AI thrives in structured, reliable data environments, while poor‑quality data leads to bias, errors and unreliable decision‑making.
The data‑AI flywheel continuously improves outcomes, with better data enhancing AI performance and AI actively enriching data systems.
Agentic AI preparedness requires unified data foundations, metadata‑driven design and continuous data quality and governance practices across the enterprise.
AI agents not only consume data, but continuously improve it, creating a self reinforcing cycle where better data drives smarter AI outcomes.
Alexy Thomas
Partner, Technology Consulting, EY India
For your convenience, a full text transcript of this podcast is available on the link below:
Welcome to EY.ai Unplugged – Season 2: AI Agents of Impact. I am Pallavi, your host for today. In this series, we speak with leaders who are shaping the future of intelligent, autonomous systems and redefining how AI drives business transformation.
In today’s episode, we are joined by Alexy Thomas, Partner, Technology Consulting at EY India. Alexy brings deep expertise in data, AI-led transformation and enterprise-scale technology solutions. He sheds light on how organizations can prepare for the next evolution of AI Agentic systems powered by trusted and high-quality data.
Thank you, Alexy, for joining us today.
Alexy
Thank you.
Pallavi
Why is high-quality, trusted data so critical for enabling Agentic AI? What are the risks of relying on poor-quality data?
Alexy
High-quality data is the single biggest determinant of how effectively Agentic AI performs. These systems learn autonomously, make decisions and take actions. So, the integrity of data directly shapes the intelligence of the agent. With inaccurate, biased or incomplete data, autonomous systems can drift, produce erroneous recommendations or reinforce existing biases. But when data is trusted, contextual and continuously refreshed, AI agents can operate responsibly, make precise decisions, and evolve in line with the business goals.
In many ways, data becomes the environment an agent lives in. If the environment is noisy or confusing or misleading, the agent will struggle. If it is structured and reliable, the agent thrives.
Pallavi
As AI becomes both a consumer and curator of data, how does ‘Data Next’ support this dual role and prepare organizations for autonomous, next-generation AI systems?
Alexy
‘Data Next’ is our internal accelerator that we have built on the principle that AI and data now evolve together. On one side, it enables organizations to have AI ready data, which is clean, structured, governed, and rich with context. That is the foundation which Agentic AI systems need to make accurate, responsible decisions. On the other hand, Data Next uses AI to improve itself.
The data itself is continually improving, and it is creating what I call ‘data and AI flywheel’. AI agents classify information, detect quality issues, enrich metadata, and create a continuously improving data environment. In other words, AI is not just consuming data, it is actively curating and elevating it. When these two come together, enterprises get a self-reinforcing cycle. Better data makes the AI smarter and smarter AI keeps making the data better. That is the data and AI flywheel.
That is what ultimately prepares organizations for true autonomy systems that are reliable, scalable, and constantly learning from the data they operate in. In essence, it helps organizations map every data point that is generated or consumed. That is the vision and mission of data.
Pallavi
What are the characteristics of good quality data for training and scaling Agentic AI?
Alexy
At its core, high quality data has four defining attributes. First is accuracy. It reflects the real world as precisely as possible. Second is completeness. Key attributes are available and consistent across sources. Third is context richness. Data carries meaning, relationships, and business relevance. Fourth is timeliness. It is updated frequently enough for agents to act confidently, and fifth is fairness and bias-free design.
Data is representative, diverse, and checked for skewing, so autonomous systems do not learn or reinforce from unintended biases. But for Agentic AI, there is an additional dimension, which is key. It is interoperability. Agents interact with multiple systems. Hence, data must be structured in a way that ensures smooth, autonomous operations across environments. When these characteristics come together, the result is data that is not only AI-ready, but also AI enhancing.
Pallavi
How does effective data preparation translate into measurable business impact for Agentic AI?
Alexy
Every step of data preparation process is unlocking meaningful value for the organization. So let me bring this to life, No. 1 is automation – agents can execute tasks without manual intervention. No. 2 is higher accuracy – decisions become more precise by reducing rework and errors. No. 3 is reduced operational cost – clean, reliable data minimizes exception handling. No. 4 is improved customer experiences – agents deliver faster, hyper-personalized responses. Fifth is better governance – with complete lineage and traceability, trust increases across the enterprise.
Pallavi
What are the practical steps of leading practices that you could share with our listeners that organizations must follow to ensure they are ready to power intelligent, autonomous systems?
Alexy
There are four essential practices:
Build a unified data foundation. Break down silos and establish a single source of truth.
Investing ongoing data quality and observability. Treat data health as a continuous program, not a one-time activity.
Adopt a metadata-first design, so AI agents understand lineage, rules and context.
Enable AI-driven curation. Use machine learning and generative AI to enrich, classify, and monitor data at scale.
These steps create an environment that Agentic AI needs to function responsibly, efficiently, and autonomously, or as I call it, ‘Creator, query-able enterprise’.
Pallavi
Thank you, Alexy. Now that brings us to the end of this episode of EY.ai Unplugged Season 2. Thank you so much once again for joining us and sharing all the valuable insights on how trusted, high-quality data forms a backbone for agent care, and what organizations must do today to build for an autonomous tomorrow.
Alexy
Thank you, Pallavi.
Pallavi
To all our listeners, if you have enjoyed this episode, do follow the series for more conversations with experts who are shaping the future of AI-driven impact. Thank you for listening. Until next time, this is Pallavi, signing off.
