As government agencies and banks in India grapple with fragmented data and the rapid, often uncontrolled adoption of AI tools, SAS is positioning its decades-old governance playbook as the foundation for the next wave of agentic AI deployments.
For a company that has spent roughly five decades in data and analytics, SAS’s strategy in India is built less around chasing the newest AI trend and more around a familiar strength: getting a customer’s data, risk, and compliance house in order before layering intelligence on top of it.
That was the central theme when Noshin Kagalwalla, Vice President – Public Sector (Asia Pacific) and Managing Director, India at SAS, sat down with Express Computer to discuss the company’s momentum, and its evolving AI strategy.
From Analytics Lifecycle to Agentic AI
SAS describes its core mission as managing the entire analytical lifecycle for customers — from data acquisition and management through to analytics and dissemination — regardless of whether the underlying technology stack is SAS, open source, or a mix of both. Layered on top of this lifecycle are industry-specific solutions targeting banking, insurance, public sector, and health and life sciences.
In recent years, the company has rebuilt this stack to be cloud-native. Its Viya 4 platform is designed for multi-cloud deployment — on Microsoft Azure, AWS, or on-premises at a customer’s own infrastructure — giving organisations flexibility as they migrate workloads that were traditionally run on-premise.
Generative AI and agentic AI are now being woven into that stack across every industry vertical SAS serves, Kagalwalla said, whether that means risk management and fraud detection in banking, IFRS 17-related risk solutions in insurance, or tax and compliance platforms for governments.
However, SAS’s approach to agentic AI remains deliberately conservative for now. “Most of our agentic [capability] is right now with humans in the loop to a large extent,” Kagalwalla said, describing a philosophy of keeping human oversight embedded in decisioning workflows even as the underlying technology increasingly allows for fuller automation.
Governance as the New Table Stakes
If there is one theme SAS is leaning into hardest, it is governance. With enterprises experimenting rapidly with AI tools — often without central oversight — the company sees a growing market opportunity in what Kagalwalla called “control tower” capabilities: traceability, auditability, and governance layered across an organisation’s AI usage.
Kagalwalla pointed to the rise of “shadow AI” as a driver of this demand. As employees across organisations adopt their own AI tools independently, the resulting lack of visibility creates risk — prompting interest in a single governance layer that can track and audit AI usage enterprise-wide, delivered by a provider with a long track record in regulated industries.
Kagalwalla noted that governance conversations are increasingly happening by default rather than as an afterthought, even though customers still lead with a specific business problem they want solved. “What the customers are interested in more is what business problem are you solving for us,” he said, with accountability and auditability now folding naturally into that discussion rather than being a separate pitch. Demand for SAS’s newer AI Navigator offering, aimed at this governance layer, is still nascent — some engagements are in pilot stage — but Kagalwalla described interest as real and growing, particularly as governance shifts from a “good to have” to what she called table-stakes functionality, not just in banking but across sectors.
India’s public sector has historically been slower to adopt AI at scale, and SAS’s own government engagements reflect that caution. In Maharashtra, SAS is working with the state on sales tax and GST-related analytics, as well as building “golden records” for citizens — unified data profiles intended to support social welfare disbursement while reducing leakage.
Neither project currently uses agentic AI, Kagalwalla said, though he expects that to be a natural next step as the underlying data and analytics foundation matures.
That foundation-first approach extends to a newly announced project with India’s Regional Transport Office (RTO) system, aimed at automating operations by consolidating data from a wide range of sources. Kagalwalla drew a direct comparison to the Maharashtra GST work, where SAS pulls together data from 21 separate sources to create a single, harmonized view — a challenge he said most government departments continue to struggle with regardless of domain.
“Having multiple data sets, harmonizing that, making sure that there is a one single view” is where most government departments struggle, he said, and it’s the same underlying problem whether the use case is taxation, citizen benefits, or vehicle registration. Solving that data-harmonization problem, in her view, is a prerequisite for any advanced analytics — agentic or otherwise — that follows.
Where the Growth Is
Looking ahead over the next six to twelve months, Kagalwalla pointed to two areas of strongest momentum: risk management in banking, and real-time fraud management, particularly application fraud. Both areas are also seeing a shift in deployment model, with customers increasingly comfortable moving analytics workloads that were traditionally on-premise onto SAS-managed cloud environments.
On the public sector side, Kagalwalla emphasized the replicability of SAS’s government work. Projects like the Maharashtra GST implementation or property tax analytics work with Mumbai’s municipal body (MCGM/BMC) demonstrate a model that can scale — whether the payoff for government customers comes through improved tax compliance revenue or through plugging leakages in citizen benefit disbursement. “The scale is massive,” he said, pointing to municipal-level examples as proof that even localized deployments can generate substantial value before being extended to other states or departments.
SAS’s India strategy suggests a company betting that the winners in agentic AI adoption — especially in regulated industries and government — will be those who solve the unglamorous problems of data harmonization, governance, and auditability first.