ServiceNow, a global leader in digital workflow automation, powers mission-critical operations for most Fortune 500 organizations. The company’s cloud-based platform integrates AI to streamline processes, enhance decision-making, and deliver measurable business value.

In the first part of this three-part interview series, Vijay Kotu, ServiceNow Chief Analytics Officer, spoke with EY TMT AI Leader Vamsi Duvvuri about five key metrics for evaluating AI use cases, translating purpose into measurable value, and embedding governance into AI initiatives.

In this second installment, Kotu dives deeper into the evolving landscape of AI integration, data strategies, and the shift toward agentic AI. This segment focuses on creating business value through seamless data ecosystems and impactful AI use cases.

Integrating AI seamlessly into enterprise ecosystems

For modern enterprises, AI doesn’t exist in isolation; rather, it thrives within interconnected platforms and systems. Kotu explains that ServiceNow’s “AI Control Tower” acts as a hub for this integration, aggregating data from multiple applications and leveraging models developed across various hyperscaler environments.

By connecting with enterprise data platforms like Snowflake and Databricks, the AI Control Tower enables leaders to compute specific business cases and measure the impact of AI use cases on key business metrics.

The promise of diverse data

Data remains the “fuel” powering AI, but Kotu emphasizes that the way organizations manage and leverage it is rapidly evolving. While data silos remain a reality, enterprises must rethink strategies to unlock value, he says.

Kotu emphasizes a universal truth for modern enterprises: data will always reside in many different places, and AI relies on a diverse set of data to deliver value. As foundational large language model (LLM) capabilities become increasingly commoditized, he notes that the real differentiators will be data quality and the experience layer — the mechanisms through which AI is embedded into everyday workflows.

On the data side, Kotu explains, traditional AI models operate in a deterministic way. In these cases, diverse data sets can be combined through APIs, integrations, or curated feature sets enhanced with AI algorithms. However, agentic AI introduces a fundamental shift. Because agentic AI operates through non-deterministic actions, it requires equally flexible, non-deterministic methods for accessing and integrating data from multiple sources.

To meet this need, ServiceNow leverages its workflow data fabric and a zero–data copy approach. Instead of moving physical data, the platform creates virtual representations that allow seamless access to diverse datasets wherever they reside. According to Kotu, this architecture ensures AI can work with richer, more varied data, ultimately driving better performance and signaling the direction the industry is heading.

Top 5 AI use case categories driving the most impact

When it comes to delivering business value through AI, Kotu identifies five core categories where enterprises are seeing the most success:

1. Search, conversations, and requests

Conversational AI is transforming how employees and customers interact with systems, enabling seamless problem-solving: “I have a problem; I need to requisition something, and I’m able to self-serve just through conversation and search.”

2. Task automation and augmentation

Breaking down roles into individual tasks allows enterprises to automate specific functions, increasing efficiency and easing change management.

3. Process automation

AI enables global optimization by orchestrating workflows across multiple teams, tasks, and even departments, improving overall process efficiency.

4. Developer-focused innovation

AI is reshaping how products are designed and built, shifting ownership beyond data scientists: “Developers are creating AI, and users are creating AI too. Who developed this AI capability? It can’t be attributed to just one team, so it’s going to be a collaborative effort.”

5. Decision support

AI-driven insights are helping leaders make better, faster decisions by augmenting managerial capabilities with predictive and prescriptive analytics.

Agentic AI: The next frontier

A key shift Kotu highlights is the transition from generative AI to agentic AI. Unlike traditional models that execute deterministic tasks like forecasting or content generation, agentic AI orchestrates multiple tasks non-deterministically to achieve end-to-end process optimization: “Agentic AI is slightly different because it brings those tasks together and orchestrates them in a non-deterministic way, allowing it to complete a process.”

Kotu advises companies to start small by focusing on existing structured processes rather than reinventing workflows from scratch. For example, ServiceNow successfully applied agentic AI to customer support processes: “Customer support is one of the processes we run, particularly low-complexity, high-volume tasks, using agentic AI, and we are starting to see good results.”

This approach, Kotu notes, allows organizations to drive local optimization first before moving toward broader process redesign.

“We should start talking about local optimization before coming up with brand new processes that never existed in an organization,” he concludes.

CDO Magazine appreciates Vijay Kotu for sharing his insights with our global community.