January 28th, 2026: HYSEA and IIIT-H hosted another thought-provoking edition of the TechForward Research Seminar Series focused on Data and Domain LLMs.
Anchored in our Academia-Industry collaboration model, the session brought together deep academic insight and enterprise-grade AI realities.

We were privileged to have:
• Prawaal, PhD, Principal Data Scientist, Infosys
• Manish Shrivastava , IIIT-Hyderabad
• Subrat Mohapatra, Vice President, Global Delivery Head – Data & Analytics, Infosys (Seminar Chair)

The conversation went well beyond “LLMs are powerful” to a far more important question:
How do enterprises actually make them work?
A strong consensus emerged that while large language models have remarkable reasoning and linguistic capabilities, their real value depends on how well they are grounded in enterprise data, domain context, and business workflows.

The discussion highlighted the growing need to move away from zero-shot, general-purpose models towards specialized, domain-adapted systems that are trained, grounded, and evaluated for real-world use cases.
Key takeaways included:
• Data is the true differentiator, quality, structure, and relevance matter more than sheer model size
• RAG and domain grounding are essential to make models enterprise-ready
• Smaller, specialized models often outperform large generalist models in accuracy and reliability
• Evaluation frameworks need to reflect real business outcomes, not just benchmark scores
• Enterprises today focus on cost reduction and risk mitigation, even more than pure productivity gains

Real-world use cases were discussed, from insurance claims and fraud detection to risk assessment, underwriting, and review summarization, including how companies are already using LLMs to create tangible business value.

The session also focused on realities around compute costs, human-in-the-loop validation, regulatory exposure, and geopolitical influences on foundational model development – all critical considerations as enterprises scale AI adoption.
TechForward helps cut through hype and bring grounded, research-driven, practitioner-led insight to Hyderabad’s technology leadership community.

Thank you to our speakers, IIIT-Hyderabad, and the Infosys team for making this an exceptional evening of learning and dialogue.
The future of AI is not just bigger models, it is smarter, domain-aware, enterprise-ready intelligence.