TLDR: Ashvin Vellody, Partner at Deloitte India, highlights that agentic AI is revolutionizing business models for IT firms by enabling autonomous, adaptive systems that solve entire workflows rather than isolated tasks. This shift promises faster execution, outcome-based engagements, and significant efficiency gains, despite challenges in data, integration, security, and change management. Deloitte’s reports indicate a strong adoption trend in India, with over 80% of organizations exploring AI-driven autonomous agents.
Bengaluru – Agentic Artificial Intelligence (AI) is poised to fundamentally transform the landscape for IT services firms, unlocking new business models and driving unprecedented operational efficiencies, according to Ashvin Vellody, Partner at Deloitte India. In a recent interaction, Vellody emphasized that agentic AI moves beyond traditional automation and even Generative AI (GenAI) by enabling systems to plan, reason, and act autonomously across entire workflows.
Unlike traditional automation, which focuses on repetitive tasks, or GenAI, which excels at content generation and summarization, agentic AI is designed to solve complex business problems by orchestrating multiple specialized agents. “The real opportunity lies in solving classes of problems by ‘agentising’ entire process flows,” Vellody explained. This involves seamlessly stitching together various single-purpose agents that collaborate to deliver comprehensive outcomes across a workflow. For instance, in a customer onboarding process, different agents might handle document validation, system updates, and follow-up triggers, collectively ensuring a complete and efficient outcome.
For IT firms, this paradigm shift translates into new business models centered on IP-led, product-driven services. It also redefines how services are priced, moving from traditional effort-based models to outcome-based engagements. The delivery model is also transformed, fostering a collaborative environment where humans and AI agents work in tandem. These agents are credited with accelerating speed-to-market, reducing duplication, and potentially improving win rates by up to 30%. Practical applications are already emerging; a global manufacturer, for example, has deployed agents to detect shipment distress and autonomously manage support tickets, while a consumer enterprise has streamlined procurement through multi-agent orchestration.
However, the implementation of agentic AI is not without its hurdles. Vellody identified several key challenges. Firstly, data readiness is critical, as many organizations lack the clean, structured, and accessible data necessary to support agent workflows. Secondly, integration poses a significant barrier due to prevalent legacy systems, fragmented APIs, and siloed platforms, which hinder end-to-end agent operations. Thirdly, robust security and governance frameworks are essential to mitigate risks such as errors, hallucinations, and compliance issues, given the autonomous nature of these agents. Beyond technical aspects, a “softer challenge” lies in change management, as business teams may be hesitant to adopt agentic AI due to a lack of immediate visible value and the disruptive impact on existing roles.
To accelerate outcomes and compress time-to-value across functions like marketing, finance, and customer service, agentic AI enables businesses to process information and make decisions more rapidly. By deploying agents across a process, tasks can be handled continuously and in parallel, eliminating manual bottlenecks. In finance, agents can automate reconciliations, reviews, and reporting, while in marketing, they can streamline campaign quality assurance and performance tracking. A global consumer company successfully used agentic AI to optimize procurement through multi-agent collaboration across sourcing, approvals, and vendor management. The key to success, Vellody noted, is building a central library of reusable agents and establishing cross-functional teams for rapid prototyping and deployment, creating a “multiplier effect” across the business.
Addressing the risk of fragmentation and bottlenecks during integration, Vellody stressed the importance of a central design authority to guide the building, deployment, and management of agents. This includes standardizing tools, platforms, and DevOps practices to ensure interoperability and consistent performance. Without such governance, agents developed in silos can lead to duplication, integration problems, and conflicting actions. A centralized platform and a library of reusable agents helped one global consumer company avoid these pitfalls. Furthermore, a “value hub” is crucial for prioritizing high-impact use cases and tracking ROI, preventing teams from pursuing low-value initiatives. Treating agentic AI as a horizontal capability with clear governance and shared infrastructure is paramount for solving business-critical problems.
Also Read:
- McKinsey’s Insights: Six Key Lessons from Over 50 Agentic AI Deployments
- Agentic AI Transforms Network Operations with Autonomous Self-Healing Capabilities
Deloitte’s “Fourth wave of the state of Generative AI in enterprises report (India insights)” reveals that over 80% of surveyed organizations in India are actively developing AI-driven autonomous agents. This trend is bolstered by India’s accelerated digitization, with initiatives like UPI-enabled payment infrastructure and UIDAI creating a fertile ground for agentic AI adoption.
