AI Agents, Unified Data, and the Road to Real ROI
In an exclusive interview with Express Computer, Natalie Mead, Vice President – Solutions Engineering, APJ at Snowflake, discusses how enterprises across India and APJ are moving beyond AI as a trend to real-world application. She talks about the growing role of agentic AI, and how Snowflake Cortex Agents are helping businesses automate workflows, unlock insights from both structured and unstructured data, and build conversational intelligence with enterprise-grade governance.
Natalie underlines the critical need for a strong data foundation, stating that there can be “no AI strategy without a data strategy”. With Snowflake’s unified AI Data Cloud architecture, organisations can eliminate data silos, integrate large language models securely, and rapidly innovate while maintaining compliance and performance at scale.
Agentic AI is being hailed as a game-changer for enterprise automation. How is Snowflake enabling businesses to leverage this technology for intelligent, data-driven decision-making?
We believe that AI agents will be essential to the enterprise workforce, enhancing productivity for teams across customer support, field technicians, analytics, engineering, and more, freeing up valuable employee time to focus on higher-value business challenges. AI agents are moving beyond basic automation, dynamically handling multi-step actions and reasoning.
For AI agents to work at scale, they need secure connections to enterprise data and unified governance to manage their access, similar to existing controls that enterprises need for their teams. They must also follow data policies, access multiple sources efficiently, and retrieve accurate information to deliver reliable, high-value outcomes.
Snowflake Cortex Agents enable businesses to build AI-driven applications that can use both structured and unstructured data. By using Cortex Search, Cortex Analyst, and LLMs, these agents can break down complex queries, retrieve relevant data, and generate precise answers.
With Snowflake Cortex Agents, enterprises are driving operational efficiency and cost reduction while enabling data-driven decision-making and faster time to market. This ultimately leads to enhanced customer satisfaction through personalised service and empowers workforce innovation by shifting focus from routine tasks to strategic initiatives.
Snowflake Cortex Agents are at the forefront of the Agentic AI shift. Can you share how enterprises across India and APJ are using them to build autonomous, real-time AI applications—and any real-world examples where these deployments have driven business impact or opened new growth avenues?
Cortex Agents represent a major step forward in enterprise conversational intelligence, combining the strengths of Snowflake’s Cortex Search and Cortex Analyst into one unified solution. These agents are capable of orchestrating across both structured and unstructured data sources, allowing them to plan tasks, utilise the appropriate tools to execute them, and generate well-informed responses. This enables organisations to manage complex, multi-turn conversations efficiently while maintaining high standards of security and governance. Across India and the APJ region, we are seeing businesses use Cortex Agents to great effect in several key areas. In finance, for instance, companies are automating financial forecasting by integrating real-time revenue data with market analysis to enable faster and more accurate decision-making. In operations, organisations are reducing delays and cutting costs by analysing inventory levels alongside logistics contracts. Engineering teams are leveraging the technology to identify trends in bug reports and customer support transcripts to proactively enhance product quality. Marketing departments are using it to personalise campaign targeting through the analysis of customer behaviour and ad imagery. Even in human resources, AI agents are helping employees compare healthcare plans by considering both policy documents and individual salary data. These examples show that Cortex Agents are already delivering significant value, enabling enterprises to extract meaningful insights quickly and apply them to real business challenges.
In a landscape where data is often fragmented, how does Snowflake’s unified data architecture help break silos and accelerate AI/ML innovation across the enterprise?
There is no AI strategy without a data strategy. Data is the foundation of AI.
Yet, many enterprises remain ill-equipped to fully leverage their data, with information fragmented across disparate systems. Furthermore, enterprise AI must be reliable, secure, and, most importantly, built on the organisation’s governed data foundation. AI is only as effective as the quality of the data it relies on. Deploying AI without clean, centralised, and well-structured data is unlikely to succeed or deliver the desired outcomes. In today’s landscape, where generative AI is driving an explosion in data creation, organisations need a unified, governed architecture that allows every team member to securely access and activate data without relying on technical resources.
Snowflake’s single platform eliminates data silos and simplifies architectures. Snowflake’s AI Data Cloud revolutionises enterprise AI/ML innovation by providing a comprehensive platform that consolidates data warehousing, data lake, and data sharing capabilities into a single system. This integration eliminates traditional data silos by enabling seamless access to data across departments while maintaining robust governance and security controls. The architecture’s elastic computing capabilities, combined with native support for various data types and real-time access, allow organisations to efficiently scale AI/ML workloads and train models using the most current information. Snowflake AI Data Cloud empowers enterprises to collaborate with data, build data applications, and innovate faster, making it the easiest, most cost-effective, and flexible enterprise data platform for customers.
Large Language Models are becoming essential in enterprise AI strategies. How is Snowflake helping organisations seamlessly integrate LLMs and other AI models into their data operations?
We continue to provide enterprises with the data foundation and cutting-edge AI building blocks they need to create powerful AI and machine learning apps with their enterprise data. For example, when accessed in Snowflake Cortex, Snowflake Arctic, the most open, enterprise-grade LLM will accelerate customers’ ability to build production-grade AI apps at scale, within the security and governance perimeter of the AI Data Cloud.
We also ensure customers have access to industry-leading LLMs by integrating them directly into the Snowflake Cortex platform. This allows businesses to access these models within their data environment without needing to set up or manage infrastructure. Snowflake Cortex gives enterprises instant access to industry-leading large language models (LLMs) trained by researchers at companies like Anthropic, Mistral, Reka, DeepSeek, Meta, and Google, including Snowflake Arctic, an open enterprise-grade model developed by Snowflake. We currently have 19 LLMs on our platform. Because these LLMs are fully hosted and managed by Snowflake, using them requires no setup. This integrated approach helps organisations accelerate their AI initiatives while maintaining data security, access to scalable computing resources, governance, and operational efficiency. At Snowflake, we continue to prioritise giving customers access to the newest and most powerful LLMs in Snowflake’s AI Data Cloud.
As India moves from ‘AI enthusiasm’ to deeper maturity, what advice would you give to tech leaders looking to build a scalable and future-ready AI infrastructure?
2025 will be defined by a shift from AI “hype” to AI utility. Organisations have been rapidly adopting AI to stay competitive and seize new opportunities. However, to truly benefit from this technology, the conversation must shift from whether an organisation uses AI to the value it aims to achieve with it. Enterprises must start to identify their goals for AI adoption, whether that be getting the information they need faster, accelerating strategic decision-making, speeding up productivity, or something else.
While not every application will be AI-powered, those that incorporate language models, knowledge repositories, and human input will evolve and improve over time. The key to achieving real ROI is to keep the focus on applying AI in places where it delivers real bottom-line impact, and follows an appropriate level of checks and balances with human input and reliable data sources.
