How Birlasoft is Solving GenAI’s ‘Last Mile’: Exclusive Interview with Dr. Selvakumaran Mannappan ,COO, Birlasoft

As cloud adoption matures, the competitive edge now lies in how intelligently and purposefully it is architected to serve specific business domains. 

The focus today is not just on where workloads are hosted, but how cloud-native, multi-cloud strategies can drive innovation, agility, and differentiation at scale - whether enhancing customer experience, ensuring regulatory compliance, or accelerating innovation. This requires cloud-native architectures that are modular, explainable, and tailored to embed intelligence where it drives the most impact. 

With the rise of GenAI, Agentic AI, and real-time analytics, platforms must enable rapid experimentation and deployment while maintaining strong governance and control. Trust, transparency, and auditability should be built in from the start, not retrofitted later. As hybrid and multi-cloud become the norm, observability is key - not just to control costs but to maximize ROI from critical workloads. 

Ultimately, this evolution is not just technological - it’s organizational, requiring domain expertise, cross-functional collaboration, and a mindset of continuous innovation. Those who get it right will unlock the cloud’s full potential as a catalyst for transformation in a rapidly changing digital world.

Agentic AI offers powerful potential through context-aware decision-making and a level of autonomy once out of reach. But this capability must be matched with strong, precise control. At its core, agentic AI must be explainable - decisions should be transparent, auditable, and understandable within context.

Building trust and resilience goes beyond technical safeguards. Enterprises need robust governance: curated model access, clear usage policies, human-in-the-loop oversight, and continuous compliance monitoring. Architectures should include model-agnostic layers to preserve business logic even as models or APIs evolve - ensuring adaptability without disruption.

Ultimately, the true test of agentic AI lies not in its intelligence, but in the strength of its governance. Responsible adoption depends on embedding trust, accountability, and oversight from the outset - not as an afterthought.

Most GenAI pilots fall short not due to lack of ambition, but because experimentation is disconnected from execution. They often run in silos, with no clear path to production or integration into enterprise workflows. Generic models are frequently used without adapting to domain-specific data, compliance needs, or existing systems.

At Birlasoft, we solve this challenge with our homegrown Generative AI platform – Birlasoft Cogito built to operationalize GenAI at scale. Birlasoft Cogito drives enterprise transformation by automating processes, enhancing decisions, and improving performance. Its model-agnostic architecture ensures resilience and future-readiness as LLMs evolve.

What sets Birlasoft Cogito apart is its comprehensive library of accelerators, tools, and frameworks that support every stage of the GenAI journey—from exploration to deployment and scale. This enables rapid movement from pilot to production, delivering real value across areas like supply chain, customer analytics, and product innovation.

To truly scale GenAI, organizations must rethink talent models beyond traditional roles and departments. It’s not enough to train a centralized AI team - AI fluency must be democratized across the enterprise. This means blending deep technical skills with domain expertise and integrating human-machine collaboration into daily workflows.

Personalized, role-specific learning is key. A finance analyst, developer, or customer service agent each needs tailored AI training. Embedding tools like Copilot into everyday apps promotes on-the-job learning, allowing fluency to grow through real use.

Equally important is cross-functional collaboration. Instead of isolated centers of excellence, agile pods - combining business users, data scientists, product managers, and engineers—can drive faster adoption and innovation by solving real-world problems together.

In the GenAI era, a successful hybrid talent model makes AI fluency a shared, scalable capability - contextual, practical, and embedded in the flow of work.

Birlasoft is cultivating GenAI fluency through a structured, role-relevant, and culture-driven hybrid talent model, integrating enterprise-wide enablement, real-world experimentation, and embedded AI solutions across business functions.

• Enterprise-Wide GenAI Enablement with Role-Specific Champions:
  Birlasoft has trained over 90% of its workforce, including functions like HR, legal, and presales, in foundational GenAI skills, focusing not only on tools like Microsoft Copilot but also on real-world usage. Each business function has designated Copilot Champions, responsible for inspiring peers, integrating GenAI into everyday workflows, and building role-specific prompt libraries to support consistent, contextualized use of GenAI as an assistant in daily work.

• Culture of Experimentation: GenAI Day, Prompt-a-thons, and Cross-Functional Hackathons:
  To promote cross-functional collaboration and grassroots innovation, Birlasoft runs multiple high-engagement events:

• GenAI Day brings together domain experts and tech teams to share real use cases.

• Prompt-a-thons and Copilot Excitement Days enable employees to form cross-functional pods and pitch innovative ideas on how Copilot can enhance their unique work contexts, from drafting legal documents to streamlining presales operations. These formats are not just awareness exercises; they operationalize AI thinking and innovation within teams and improve adoption by making learning hands-on and team-driven.

• Embedded GenAI Solutions That Drive Business Outcomes:
  Birlasoft embeds GenAI across its ecosystem through proprietary and platform-integrated tools:

• B-Hive, a GenAI-powered virtual assistant, handles 94% of HR policy queries with personalized responses, improving employee self-service.

• Solución, integrated with ServiceNow, acts as the first responder for IT tickets across 10+ internal apps. It resolves 73% of knowledge-based issues and has cut mean-time-to-resolution from 55 hours to near-zero.

• The Birlasoft Cogito platform, built on a model-agnostic architecture, accelerates the development and deployment of domain-specific GenAI solutions across finance, supply chain, and manufacturing.

• Hybrid Talent Model Fueling Cross-Functional AI Pods:
  Birlasoft’s hybrid talent model combines internal upskilling with AI-driven skill platforms to create cross-functional AI pods, teams that bring together business and tech talent to co-develop solutions. This ensures GenAI tools are not only technically sound but also deeply aligned to business context, unlocking value at the intersection of domain knowledge and digital capability.

• GenAI Centre of Excellence and Use Case Factory:
  Birlasoft’s GenAI Centre of Excellence, built in partnership with Microsoft, has delivered over 50 enterprise use cases and continues to expand the company’s innovation pipeline. The CoE acts as both a skill hub and an incubator, feeding back learnings into the enterprise-wide enablement programs to sustain momentum.

Together, these initiatives reflect Birlasoft’s commitment to building a GenAI-ready workforce not through generic upskilling but by embedding AI into roles, tools, culture, and business outcomes.

AI transformation in 2025 demands a new kind of leadership - focused, adaptive, and deeply responsible.

First, clarity on outcomes is non-negotiable. While pilots and prototypes have their place, true enterprise value comes from measurable impact - whether in efficiency, revenue, or customer experience. Every AI initiative must be tied to a clear business case aligned with strategy and scale.

Second, cultural readiness is critical. AI is as much a mindset shift as a tech shift. Leaders must guide teams through uncertainty, encourage experimentation, and eliminate fear. This means investing in storytelling, skill-building, and empowering champions to drive change.

Third, trust and responsibility must be built in from day one. From data privacy to bias and cybersecurity, AI must be explainable, auditable, and governed to ensure safe, ethical use.

In a world defined by complexity and rapid change, clear outcomes, cultural leadership, and responsible innovation are the cornerstones of successful AI transformation.