Enterprise AI is driving business transformation, with over 50% of AI use cases delivering impact, according to Infosys. Success is highest in professional services and tech, while financial services face regulatory hurdles. Satish H C, EVP, chief delivery officer, Infosys, tells Padmini Dhruvaraj, that the company is investing in small language models, Agentic AI, and workforce transformation to enhance AI-driven automation. Excerpts:
According to Infosys study only 20% of AI (artificial intelligence) use cases fully meet business goals, with 31% nearing success. What key challenges hinder AI deployments?
Achieving transformative change requires recognising complexity and organisational readiness. While incremental value is possible, true transformation comes from re-imagining processes with an AI-first approach, not just augmenting existing ones. This shift depends on workforce engagement, change management, training, and budget support. Expectations also vary by AI type. Generative AI excels in chatbots, question and answer, and document summarisation, while AI for business value relies on traditional models, BI queries, or statistical methods for accurate predictions. Achieving this requires agentic frameworks, interconnected models, and significant investment.
How is Infosys helping clients optimise AI investments to ensure sustainable cost reductions without compromising performance?
We are helping clients optimise AI investments to ensure sustainable cost reductions without compromising performance through several strategic approaches. To name a few – AI tooling leveraging AI foundry, AI factory, Agentic AI for enterprise transformation, product centric operating model and setting up value office. AI foundry and AI factory to experiment and incubate new technologies, develop new patterns etc. The AI factory then turns these learnings into products and platforms. Together, they form the hub of AI innovation, experimentation, building and training customer models and then eventually leading to value realisation, workforce engagement, and governance at scale. With our Responsible AI 3S framework (Scan, Shield, Steer), we are ensuring innovation is safe for adoption for our customers. Agentic AI for enterprise transformation involves creating multiple small AI agents, each playing a specific role or carrying out a specific task including taking decisions, plan actions, and learn from experiences, enabling end-to-end processes.
Are there any Infosys-led AI implementations that have significantly transformed business outcomes for clients both in terms of efficiencies and return on investment?
In the US CPG (Consumer Packaged Goods) sector, our Media Mix Optimiser slashed implementation time significantly and boosted accuracy by 10-15%, optimising marketing budgets and maximising ROI (Return on Investment) through automated, data-driven decision-making. Another is for the work we did for a US electric utility, our Enterprise Analytical Platform reduced data processing time, improved forecasting accuracy, and cut operational costs, enhancing asset planning and monitoring.
Can you share any examples of companies that have successfully adopted agentic AI models for enterprise transformation?
While most transformation projects will involve implementing Agentic AI framework, our largest agentic AI implementation transformed the workflow of several inventory analysts and users. By leveraging generative AI and a pretrained large language model (LLM), we automated data insights, saving analysts over 10% of their time. In another example, an American tech conglomerate’s support team faced high ticket volumes, siloed information, and slow resolutions, impacting customer experience. Infosys’ Agentic AI-powered assistant and chatbot streamlined support by integrating multiple knowledge sources, enabling faster resolutions, doubling first-day issue resolution, and boosting agent productivity.