As enterprises push AI adoption beyond pilots, the real test lies in operationalizing intelligence within governed, auditable workflows rather than treating it as a standalone experiment. In this conversation with CIO&Leader, Sunil Pandita, SVP and Head of Business, India & South Asia at Newgen Software, unpacks what it takes to move decision intelligence from fragmented systems to enterprise-wide execution. He discusses why Agentic AI must be embedded within existing orchestration frameworks rather than operating autonomously, the guardrails BFSI and government bodies need before trusting AI with underwriting, lending, and claims decisions, and why explainability and human oversight remain non-negotiable in regulated environments. Pandita also outlines the technical capabilities—from data unification to model governance—that will distinguish enterprises that successfully scale AI from those stuck in pilots.
SVP and Head of Business, India & South Asia
Newgen Software
CIO&Leader: How can enterprises overcome fragmented architectures to enable enterprise-wide decision intelligence?
Sunil Pandita: Many enterprises have successfully digitized individual functions, yet decision-making often remains fragmented across systems, workflows, content repositories, and teams. While AI adoption is gaining momentum, the Economic Survey notes that only 21% of Indian banks and financial institutions have begun implementing or building AI solutions for core operations, highlighting that the industry is still in the early stages of operationalizing intelligence at scale. Overcoming this challenge requires a unifying operational layer that connects data, processes, AI, and human decisions within a shared execution context. Decision intelligence emerges when information flows seamlessly across the enterprise, rather than being trapped within isolated systems. In banking, in particular, intelligence must operate within workflows, not outside them. By integrating front- and back-office processes, contextual information, and governance mechanisms, organizations can move beyond isolated automation and enable coordinated, real-time decision-making that is scalable, auditable, and aligned to business outcomes. Ultimately, enterprises that succeed will be those that build the operational agility to anticipate change rather than merely respond to it, staying ahead of what’s ahead.
CIO&Leader: How is Newgen defining and operationalizing “Agentic AI” within enterprise workflows, and how does it differ from the automation capabilities of traditional low-code/BPM platforms?
Sunil Pandita: Traditional automation focuses on executing predefined rules and workflows efficiently. Agentic AI introduces a new dimension by enabling systems to reason, interpret context, recommend actions, and dynamically adapt within established business boundaries. The real value, however, does not come from autonomous intelligence operating in isolation. It comes from embedding intelligence directly into governed enterprise workflows. At Newgen, we view Agentic AI as an extension of enterprise orchestration, where AI agents, systems, data, and human decision-makers operate within a shared execution framework. This allows organizations to move beyond task automation toward contextual, adaptive, and auditable decision-making while maintaining the governance, transparency, and accountability required for business-critical operations.
CIO&Leader: As agentic systems begin making autonomous decisions within business processes, what governance and human-in-the-loop mechanisms are enterprises building to maintain control and auditability?
Sunil Pandita: Human oversight remains critical in regulated environments where decisions have significant financial and compliance implications. AI can enhance efficiency, surface insights, and automate routine activities, but accountability cannot be delegated entirely to algorithms. In lending, compliance, and customer service, organizations must ensure that AI operates within governed, auditable workflows, with humans providing judgment, exception handling, and oversight. The most effective model is one where AI amplifies human expertise rather than replacing it. This approach strengthens trust, improves transparency, and ensures that institutions can confidently explain and defend decisions to regulators, auditors, customers, and other stakeholders.
CIO&Leader: How is the architecture of low-code platforms evolving to support the integration of LLMs and AI agents without compromising security, scalability, or compliance?
Sunil Pandita: As AI adoption moves from experimentation to enterprise-wide deployment, organizations require platforms that can integrate intelligence without disrupting governance or operational control. Modern low-code architectures are evolving to serve as orchestration layers that connect workflows, AI models, content, business rules, and enterprise systems within a unified framework. This approach enables organizations to embed AI directly into operational processes while maintaining security, auditability, and regulatory compliance. Equally important, it helps enterprises avoid creating new silos by ensuring that AI-driven decisions remain connected to business context, human oversight, and established governance policies. The objective is not simply to deploy AI, but to operationalize it responsibly at scale.
CIO&Leader: In regulated sectors like BFSI, what specific technical guardrails are needed before AI agents can be trusted to make decisions in processes such as underwriting, claims, or lending?
Sunil Pandita: Before AI agents can be trusted in decision-critical processes such as lending or underwriting, enterprises must establish clear guardrails around explainability, auditability, governance, human oversight, and policy enforcement. Trust in AI is not created by automation alone; it is created by ensuring that every decision remains transparent, traceable, and accountable. Banks must recognize that speed alone cannot be the measure of success. Credit decisions carry accountability obligations and therefore must be explainable, traceable, and auditable. The objective is not unrestricted AI autonomy but controlled autonomy, where intelligence operates within governed business processes. This becomes particularly important given that while more than 70% of enterprises are investing in GenAI, fewer than 20% have established mature governance frameworks to manage AI at scale. AI-driven underwriting can accelerate pre-screening, approvals, and risk assessment, but decisions should remain embedded within transparent workflows that provide visibility into how outcomes are reached. When governance, compliance controls, and human oversight are integrated into the decisioning process itself, banks can achieve faster lending outcomes while maintaining trust, regulatory readiness, and responsible risk management.
CIO&Leader: How does Newgen’s platform approach the challenge of unifying structured and unstructured data to make AI-driven decisioning viable across legacy enterprise systems?
Sunil Pandita: One of the biggest barriers to AI adoption is not the lack of data, but the fragmentation of data across systems, documents, communications, and business applications. AI-driven decisioning depends on the ability to integrate these diverse information sources into a common operational context. Newgen addresses this challenge by combining process automation, content services, document-centric workflows, and enterprise integration capabilities on a unified platform. This enables organizations to connect structured data from core systems with unstructured information embedded within documents, customer interactions, and business content. As a result, AI can operate with greater context, enabling more accurate, timely, and explainable decisions across complex enterprise environments.
CIO&Leader: What role do intelligent document processing and content services play in enabling AI-first operations for large, document-heavy enterprises like banks and government bodies?
Sunil Pandita: For document-intensive industries, intelligence is only as effective as the quality and accessibility of information available to it. Banks, insurers, and government organizations generate vast volumes of unstructured content that must be captured, classified, understood, and acted upon efficiently. Intelligent document processing and content services help transform documents into actionable business intelligence by extracting relevant information, organizing content, and making it available within operational workflows. This reduces manual intervention, improves process efficiency, and strengthens decision quality. More importantly, it enables AI to operate within the flow of work, where decisions are informed by complete context rather than isolated data points, creating a foundation for scalable and accountable AI adoption.
CIO&Leader: How are enterprises addressing model explainability and auditability when deploying AI within regulated, decision-critical government and BFSI workflows?
Sunil Pandita: In regulated sectors such as BFSI, explainability and auditability are becoming foundational requirements for AI adoption. Organizations must be able to understand how decisions are made, what information influenced them, and whether outcomes remain aligned with regulatory and business policies. This is particularly important as AI plays a greater role in customer-facing and risk-sensitive workflows. Rather than allowing AI to operate as a black box, leading enterprises are embedding governance, transparency, and monitoring directly into operational processes. This ensures decisions remain understandable, defensible, and auditable while preserving the trust of regulators, customers, and stakeholders.
CIO&Leader: With enterprises running a mix of legacy core systems and modern AI layers, what integration challenges are most common, and how is Newgen’s architecture designed to bridge them?
Sunil Pandita: The most common challenge is that intelligence often resides separately from execution. Enterprises have invested in core systems, digital channels, automation initiatives, and AI solutions over time, but these capabilities frequently operate in silos. This creates fragmented workflows, inconsistent decision-making, and limited operational visibility. Bridging this gap requires a unified architecture that can connect systems, data, documents, workflows, and AI-driven insights without requiring organizations to replace existing investments. Newgen’s approach focuses on creating an orchestration layer that brings these components together within a common operational framework. This enables enterprises to modernize incrementally while ensuring intelligence is embedded within business processes rather than layered on top of them.
CIO&Leader: What separates banks that successfully operationalize AI from those that remain stuck in experimentation?
Sunil Pandita: The differentiator is not access to AI models but the ability to integrate intelligence into day-to-day operations. Many institutions achieve success with pilots yet struggle to scale because AI remains disconnected from core workflows and decision processes. Successful banks treat AI as an operating capability rather than a standalone feature. They establish strong governance frameworks, embed intelligence into business processes, and connect data, systems, and human decision-makers through a unified execution model. This allows insights to translate into coordinated action. Institutions that achieve this shift move beyond isolated experimentation and create sustainable, enterprise-wide business value from AI investments.
CIO&Leader: As AI adoption scales across large enterprises, how is Newgen thinking about infrastructure choices — cloud, on-premises, or hybrid — to support performance, data residency, and cost considerations?
Sunil Pandita: Infrastructure decisions ultimately depend on business, regulatory, and operational priorities. Many large enterprises, particularly in regulated sectors, require the flexibility to balance innovation with data residency, compliance, and risk management requirements. Our perspective is that organizations should not have to choose between flexibility and control. Enterprises increasingly prefer deployment models that allow them to leverage cloud scalability where appropriate while retaining sensitive workloads within environments that align with regulatory and governance obligations. The focus should remain on enabling a secure, scalable, and future-ready operating environment in which AI, workflows, content, and enterprise systems can function seamlessly, regardless of the underlying deployment model.
CIO&Leader: Looking ahead, what technical capabilities do you believe will separate enterprises that successfully scale AI-first operations from those that remain stuck in pilot deployments?
Sunil Pandita: Leading organizations will combine intelligence, governance, and execution within a single operational framework. The technical capabilities that matter most include workflow orchestration, enterprise integration, data unification, real-time visibility, model governance, and end-to-end auditability. Equally important is the ability to embed compliance and control mechanisms directly into operational processes rather than treating them as external checkpoints. As organizations move beyond pilots, success will depend less on standalone AI models and more on their ability to connect systems, workflows, content, and decisioning layers into a coordinated execution framework that can scale with confidence and control.