New Delhi, July 18 -- India-based outsourcing firm WNS has recently made headlines following Capgemini's $3.3 billion acquisition of the firm, wherein WNS will be integrated into Capgemini's Global Business Services (GBS) unit. Founded in 1996 as a captive unit for British Airways, WNS has evolved into a global BPO company serving over 700 clients across industries and operates through 64 delivery centres in 13 countries. Though in its early days, Capgemini's acquisition of WNS signals a resurgence of the IT-plus-BPO model, driven by the need for AI integration. In an exclusive interaction with TechCircle, Gautam Singh, Head - Analytics, Data & AI at WNS, discusses the transformative potential of Agentic AI, challenges faced by tech leaders in scaling enterprise AI and the way forward. Edited excerpts.
How has enterprise AI maturity evolved in the last 1-2 years, and what are the driving forces?
Enterprise AI has matured significantly in the past year, shifting from trial projects to strategic implementations. Organisations now understand enterprise AI across three levels: data (harnessing unstructured data), transformation (leveraging Generative AI), and orchestration (using Agentic AI to optimise the partnership between human ingenuity and AI). Key drivers of this evolution include: advances in Large Language Models enabling versatile business applications; C-suite commitment to scaling AI programs; the rise of microservices for scalable solutions; increased AI democratisation via Generative AI; synthetic data generation reducing reliance on labelled data; and the emergence of Agentic AI for reimagining business processes and reinforcing learning.
How do agentic systems differ from traditional automation, and why is this important?
Traditional automation excels in structured environments with predictable inputs, executing rule-based tasks efficiently. However, it lacks adaptability and real-time reasoning. Agentic AI, conversely, combines autonomy, context-awareness, and collaboration. It interprets intent, decomposes complex goals, learns from evolving data, and orchestrates outcomes. This adaptability is crucial in environments where traditional automation falls short, representing a shift towards scaling and orchestrating intelligence. Benefits include human-like reasoning, collaborative agent networks, responsiveness to complexity, faster insights, and suitability for ambiguous scenarios.
What are the risks of increasingly independent AI agents, and how are you ensuring client safety?
Greater AI autonomy introduces risks, including misaligned objectives, ethical and compliance violations, loss of human insight, security vulnerabilities, and over-dependence on AI judgment.
Will AI agents eventually possess domain expertise across enterprises, or will they remain task-specific?
AI agents require domain context, especially in dynamic business environments. Advances in Agentic AI enable agents to learn continuously, interact with domain-specific data, and integrate with business frameworks, becoming intelligent participants in enterprise workflows. For example, a demand forecasting agent can evolve to understand seasonality, vendor behavior, and cost implications. Multi-agent systems, where each agent handles a distinct function within a domain-aware ecosystem, offer a balance of scalability, precision, and adaptability. At WNS, domain experts guide AI's contextual learning within a Human-in-the-Loop framework.
What are the biggest challenges today to scalable and trustworthy enterprise AI?
Scalability hinges on effectively integrating data systems, ensuring robust governance, and maintaining model performance. Siloed data, weak governance, and insufficient human oversight hinder AI's promise of faster decisions and intelligent automation. Addressing outdated data systems, establishing ethical guardrails, and developing skilled talent are crucial. Resilient AI systems require a cohesive ecosystem that aligns data, people, processes, and technology.
What new job roles are emerging with agentic AI, and how will this impact future job creation?
Agentic AI is transforming, not eliminating, jobs. It fosters teamwork between humans and machines, creating roles in AI behaviour design, agent management, reinforcement learning training, and AI regulation analysis. WNS has trained over 22,000 employees in areas such as Gen AI and analytics. As Agentic AI evolves, enterprises will require new skills and expertise.
What upcoming technologies are you most enthusiastic about?
Gen AI's potential remains untapped, particularly in its ability to generate data across formats and enable continuous learning through reinforcement learning. Multi-LLM orchestration platforms, with specialised AI agents collaborating on complex tasks, are another exciting frontier. Emerging agent-to-agent communication protocols will revolutionise AI interoperability. The convergence of Agentic AI, neuro-symbolic learning, and autonomous decision engines will drive enterprise intelligence from reactive analytics to predictive and prescriptive workflows. The future lies in intelligent ecosystems where AI agents and humans co-evolve, continuously learning to solve real-world problems at scale.
Published by HT Digital Content Services with permission from TechCircle.