The enterprise AI landscape has decisively shifted from experimentation to execution. For today’s data and analytics leaders, the focus is now on next-generation agentic systems that can reason, coordinate workflows and assist with decision-making across business processes. Gartner projects that 40 percent of enterprise applications will feature embedded, task-specific AI agents by the end of 2026, marking a significant leap from less than 5 percent in 2025.

This shift is accelerating at a breakneck pace, and discussions about it took center stage at CDAO UK 2026, where WNS moderated a roundtable on driving successful Agentic AI-led transformations, bringing together data, analytics and AI leaders from across industries. The candid discussion explored the foundational requirements and critical enablers required to turn Agentic AI potential into reality, delivering rapid ROI, future-ready systems and trustworthy autonomy for continuous, enterprise-wide value realization.

Shaping the Future of Enterprise Agentic AI: Insights from the Roundtable

The discussion quickly moved past the hype surrounding enterprise AI into the realities of implementation. As organizations begin exploring agentic systems for key functions, leaders are confronting a common set of questions around data readiness, governance, autonomy and safe experimentation.

Drawing on experiences from sectors such as insurance, financial services, retail, media and entertainment, participants examined what it takes to translate Agentic AI ambition into practical enterprise adoption. By the end of the conversation, the discussion had crystallized around three strategic themes, shaping how organizations are beginning to operationalize Agentic AI.

A common trap for enterprises is trying to fix the data before deciding what to do with it. Data is ever-changing, as are the requirements from that data. The room explored a complete reversal of established logic, with the idea being to start with a use case and work backward.

In this approach, enterprises would create data ponds with targeted sets of information required to solve a specific, high-priority market need – rather than wait years to build an exhaustive enterprise data lake. The ponds would be built with a shared architecture that would allow for eventual integration into a data lake, delivering faster business impact instead of relying on perfect data conditions, which may never arrive.

Data management is not a one-time project but an ongoing process that evolves with business operations. Centralized management is at odds with this reality. When responsibility for cleaning and managing data rests with a single team, issues such as duplication and errors often persist at the source, limiting the ability to keep data accurate and up-to-date.

A federated data management model was discussed as the way forward. Driven by the CXO-level leadership, this model pushes ownership down to the teams who create the data, whether in sales, invoicing or operations, ensuring the quality is fixed at the source. When the teams that input the data are accountable for its impact on the final outcome, the ecosystem remains dynamic, accurate and ready to fuel autonomous agents.

A key point recognized by leaders was that Agentic AI is not a one-size-fits-all solution for industries. The roadmap to adoption will be dictated by sector-specific regulatory pressures and risk appetites. Agentic autonomy will therefore be tailored to the enterprise context.

For instance, highly regulated sectors such as insurance may initially need tighter controls and higher human oversight, gradually re-calibrating the agent’s autonomy as it proves its reliability in data-sensitive areas such as claims and risk-sensitive areas such as underwriting. In contrast, sectors such as retail and logistics may be bolder, prioritizing speed, personalizing communication and granting agents higher autonomy to drive immediate competitive advantage. Given the rich variety of business scenarios, the choice of where and how to bring in Agentic AI depends on the nuances leaders identify.

The CXO Mandate: Driving the Enterprise Evolution

As Agentic AI adoption expands, technology alone cannot drive the shift – a point acknowledged in multiple conversations around the roundtable. Scaling autonomous systems requires a leadership-led evolution of the entire business.

Leadership Mindset as the Catalyst

The tools to scale Agentic AI are already in place. The decision to scale, therefore, does not hinge solely on technological readiness. The true catalyst lies in the vision and intent of enterprise leadership. Transformation must be top-down and CEO-led, with executive culture shaping organization ambition, speed and strategic priorities.

This mindset shapes critical early decisions, such as, should the enterprise take on complex, high-impact challenges first or build momentum through smaller, targeted pilots? The roundtable discussion made is clear that the right approach varies by sector.

What remains consistent is the need for leaders to back a clear, decisive strategy; one that embraces the realities of dynamic data and fosters a proactive, forward-looking culture.

From Operating Model to Autonomous Enterprise

Agentic AI is re-shaping the foundations of how work gets done. Instead of teams executing processes end-to-end, they are increasingly required to supervise, guide and intervene in systems that act autonomously. This shift re-defines accountability, re-shapes performance metrics and demands new approaches to building trust in decision-making.

Realizing value, therefore, depends on aligning business, data and technology teams around a shared operating model for Agentic systems. Organizations that approach this as a fundamental transformation in how work is structured, rather than a standalone technology deployment, are better positioned to move beyond isolated pilots and deliver sustained impact.

At the same time, this evolution must be carefully governed. Leaders need to balance innovation with control, using secure development sandboxes to prototype solutions and validate governance frameworks before scaling to production. This ensures that as the organization evolves, it does so within a structure of human oversight and regulatory compliance.