From Start to Scale — Transformation & Acceleration with Agentic AI

In a market defined by fierce competition and fast-moving AI innovation, CIOs are under pressure to deliver measurable business impact. Agentic AI is shifting the conversation from adoption to responsible, effective scaling. At the same time, cloud modernization is entering a new phase — one where AI agents are fundamentally changing how enterprises approach business processes and service delivery for greater agility and cost savings.

CIOs are uniquely positioned to lead this transformation but must overcome key barriers: technical debt, skills shortages and heightened security and compliance risks. How can IT leaders move beyond pilots and drive measurable ROI with agentic AI?

CIOs in our New York Community gathered recently for a Town Hall discussion to share how they are building AI-driven ecosystems to leverage agentic AI. Charlie Doubek, Global VP, Agentic AI and Cloud & Security at NTT DATA, moderated the discussion. Gartner New York CIO Community Members Sanjay Bhakta, Chief Product & Technology Officer at Conde Nast Publications, and Christopher Smith, CIO at Freshfields, served as panelists.

The discussion focused on how CIOs are tackling the technical debt, skills gaps, and cost pressures that are barriers to scaling AI agents, as well as how they are finding high-impact use cases and building the ecosystem to support them. In addition, the panelists and participants exchanged ideas about collaborating across the enterprise to leverage agentic AI for cloud modernization, innovation and measurable outcomes.

Here are the 5 key takeaways from the Town Hall discussion:

1. Managing Pressures and Expectations Around AI Adoption
   CIOs face significant pressures from stakeholders to adopt AI and to demonstrate ROI from it. One CIO noted that it is about “balancing the enthusiasm [for AI] with a regulated industry where trust and quality of service are important.” Another IT leader shared that there is a “gap between reality and expectation,” where the expectation is transformational AI and the reality is often skills gaps and legacy technology. They agreed about the need to “fix the underpinnings” of these practical challenges before realizing meaningful ROI from AI.
2. Building and Governing AI-Driven Ecosystems
   In order to successfully scale agentic AI, CIOs said that robust governance is a must –  especially to address privacy, security, and regulatory requirements. Organizations are implementing governance processes to evaluate and control tool adoption, often choosing a mix of custom and commercial solutions. Clear policies on data access, tool approval, and agent deployment are critical. One CIO said, “There is an avalanche of vendors offering tools to associates,” which requires a robust innovation governance process. Agentic AI also requires IT leaders to think about agents interacting with structured and unstructured data, and one shared they are being “very cautious on the level of autonomy of agents in this environment.”
3. Driving Enterprise-Wide Collaboration and Change Management
   Scaling agentic AI requires cross-functional collaboration, as knowledge and processes are siloed across departments. One CIO noted the critical nature of collaborating on security, privacy and regulations, saying that CIOs need to “put your Chief Privacy Officer on speed dial – and your CISO.” Change management is challenging because AI can impact job roles and responsibilities, leading to resistance. As one IT leader shared, you have to "work on the mentality of people trying new things, new processes, and new tools.” Finally, top-down leadership and clear business ownership are essential for success. Another CIO commented, “If you want an AI project to be successful, you have to have a business owner and a KPI.”
4. Practical Use Cases and Cloud Enablement
   CIOs shared that their organizations are leveraging the cloud to support AI initiatives, enabling experimentation and scalability. One shared that the cloud “allows you to try things quickly,” and others noted that it is now "table stakes – you have to be in the cloud.” They talked about use cases ranging from legal document parsing to natural language search, but mentioned that adoption is often limited by data fragmentation, regulatory hurdles, and the need for specialized models. Cloud capacity and data readiness are foundational.
   
5. Navigating the Buy vs. Build Dilemma and Vendor Proliferation
   CIOs report they are cautious about the crowded AI vendor landscape, where many offerings are just prompt engineering on standard models. The preference is often to buy rather than build, as several CIOs noted building a tool is not their core business, but evaluating real value is difficult. They recommended minimal investment in pilots and a focus on practical, immediately valuable solutions, as many tools fail to gain user adoption. One CIO said to ask yourself, “What can I get value from today without too much effort or downside?”

CIOs acknowledge there is a gap between the transformative expectations of AI and the practical challenges of legacy systems and workforce readiness. They are testing, iterating and encouraging adoption, but as one said, “Fundamentally changing business processes and mindsets is hard.” They emphasized the importance of cross-functional collaboration and change management in moving AI initiatives forward.

To collaborate with other CIOs on implementing agentic AI and other critical priorities for technology leaders, apply to join a CIO community. If you are already a community member, sign in to the app to find upcoming opportunities to get together with your peers.
 

Special thanks to NTT DATA.