Birlasoft CTO sees agentic AI reshaping enterprise operating models

Enterprises are moving beyond AI pilots, but fragmented data, weak governance and the absence of business ownership continue to impede deployment at scale. In an excvlusive interview, Ganesan Karuppanaicker, chief technology officer at Birlasoft, of Indian IT services firm, says the next phase will require companies to redesign processes around human-AI collaboration, not merely add AI tools to existing workflows.

Artificial intelligence is beginning to alter how enterprises execute work, moving from generating content and insights to making decisions and completing business processes. For IT services companies, this transition presents an opportunity but also challenges the traditional people-led delivery model. Karuppanaicker also discusses why companies remain trapped in AI pilots, how developers’ roles will change, and why governance must be enforced while AI agents are executing tasks rather than reviewed afterwards. Edited excerpts:

Enterprises are trying to move from AI pilots to production deployments. How is this changing Birlasoft’s technology and delivery strategy?

Moving to enterprise-scale AI requires much more than deploying models. It calls for new operating models, technology architectures and governance frameworks.

Our strategy rests on three priorities: AI-first process reimagination, platform-centric engineering and outcome-anchored transformation. Instead of adding AI to existing workflows, we work with clients to redesign processes and determine how people and intelligent agents can collaborate to improve speed, accuracy and business value.

Enterprises are no longer asking whether AI can be deployed. They want to know how it can improve operational efficiency, support growth or create a competitive advantage.

Our Cogito platform brings together AI orchestration, enterprise context, security and governance. When combined with our Quantum Sprint delivery model and with humans in the loop, itis delivering a two- to threefold increase in engineering throughput.

Why do so many companies remain stuck at the proof-of-concept stage?

Organisations that successfully scale AI generally perform well in three areas: foundational readiness, governance maturity and business ownership.

They invest in high-quality data, modern cloud platforms, integrated architecture and AI-ready talent. Without these foundations, AI remains limited to isolated use cases.

They also treat governance as a strategic capability. As AI moves from producing insights to executing decisions, enterprises require transparency, accountability, human oversight and clear risk controls.

Finally, AI must be treated as a business transformation programme, not simply an IT initiative. Leadership must connect the investment to defined outcomes such as productivity, customer experience, revenue growth or operating efficiency. The difference between experimentation and deployment ultimately appears in measurable results.

How will generative AI change the role of software developers?

AI will not eliminate developers, but it will elevate their role from code creators to architects, orchestrators and problem-solvers. As routine coding becomes increasingly automated, engineers will focus more on system design, AI orchestration, security, governance and aligning technology with business requirements.

The next generation of talent will need engineering skills combined with AI literacy and domain knowledge. Prompt and harness engineering, agentic workflow design, model evaluation, data fluency and responsible AI implementation will become important capabilities.

Through our Forward Deployed Engineer programme, we place AI-native talent in client environments. We are also building role-based AI fluency through an Agentic AI Academy.

Where are enterprises making the most aggressive investments to prepare for AI?

The three principal areas are data readiness, cloud optimisation and AI-enabled platforms.

Companies recognise that the quality, governance and accessibility of their data directly influence AI performance. This is driving investments in modern data platforms, governance and enterprise-wide integration.

Cloud conversations have also shifted from migration to optimisation. Enterprises are building hybrid and multi-cloud environments that can support AI workloads while balancing performance, security, regulation and cost.

There is also demand for integrated platforms connecting cloud, data, enterprise resource planning and AI. Cogito is open by design, allowing clients to bring their preferred model, development environment and cloud while applying common governance.

As AI agents begin executing business processes, can conventional governance frameworks keep pace?

Trust must be engineered into AI systems from the beginning, rather than added after deployment. Every action taken by an AI agent should be explainable, traceable and consistent with enterprise policy.

Governance also has to be applied during execution. Within Cogito, this includes policy-as-code controls at runtime, responsible-AI checks covering areas such as fairness and bias, multi-agent observability and reproducible records of agent activity.

Human oversight will remain necessary for particular decisions. The objective is to allow scale and compliance to advance together rather than treating governance as a constraint introduced after innovation.

What will be the biggest disruption to IT services over the next three to five years?

Agentic AI will be among the most transformative forces. Unlike earlier technology waves that primarily enhanced productivity or decision-making, agentic systems can reason, plan, collaborate and execute actions.

The disruption will not simply come from AI replacing people. It will come from AI reshaping operating models, reducing coordination overhead, automating decision flows and embedding intelligence into business execution.

This will change software engineering, service delivery, governance and workforce models. Enterprises will move from deploying individual AI tools towards becoming AI-native organisations built around governed collaboration between people and intelligent agents.