The enterprise technology landscape is undergoing a profound structural shift. While the initial wave of Generative AI focused heavily on boosting individual productivity, the next frontier belongs to something much more fundamental: the Agentic Enterprise.
“Most enterprises today are using GenAI to improve productivity, while leaving core operating models largely unchanged,” explains Anuj Bhalla, Senior Vice President and Global Delivery Head of Cloud & Infrastructure Services at Cognizant. “An agentic enterprise, however, places trust in AI systems to detect, decide and act autonomously within clearly defined guardrails.”
As hyperscalers inject unprecedented capital into AI infrastructure to meet the soaring demand for cloud services, traditional IT managed services are facing a forcing function. Linear, activity-based volumes are giving way to outcome-driven, self-healing architectures. In a recent conversation, Bhalla broke down what this paradigm shift looks like in practice, providing a pragmatic roadmap for leaders looking to translate raw AI investments into measurable, scalable business outcomes.
From Reactive To Predictive: The New Operating Fabric
For decades, enterprise IT operated on a reactive, human-led model. The agentic enterprise fundamentally flips this script.
“At Cognizant, we are moving from reactive, human-led execution to predictive, self-healing, machine-first operations, with humans focusing on governance, exceptions & innovation,” says Bhalla.
In practice, this means deploying autonomous AI agents that handle the heavy lifting of day-to-day operations.
“In practice, we use AI agents that autonomously triage incidents, suppress alert noise, fulfil service requests and remediate issues. Infrastructure operations are evolving in parallel, with agent-driven capacity planning, FinOps optimisation, and continuous compliance becoming embedded into day-to-day operations,” Bhalla notes. To support this shift, “Cognizant is rapidly ramping up capabilities in this space by upskilling internal teams and through strategic acquisitions like Astreya, which help supercharge autonomous IT operations capabilities and strengthen our existing foundation.”
Redefining Success Beyond The Ticket
As intelligence becomes embedded across cloud, infrastructure, and DevOps workflows, the metrics used to evaluate enterprise IT partners must change.
“Activity volume is no longer a benchmark for success. Instead, resilience, cost efficiency and business impact define outcomes,” Bhalla states. At the core of this transition is Cognizant’s Autonomous IT Operations model, which “transforms IT operations from assisted execution to autonomous delivery.” According to Bhalla, “This model is powered by a unified agentic delivery fabric, built on Cognizant’s AI Continuum. It embeds intelligence across cloud, infrastructure, and DevOps workflows, enabling predictive operations through adaptive workflows, self-healing systems, and compliant-by-design execution.”
Crucially, achieving these outcomes requires a departure from traditional legacy commercial structures. “Alongside this, Cognizant is consciously moving away from traditional fixed-cost and time-and-materials models. This transition is supported by our investments in emerging intelligent platforms, a strong strategic partner ecosystem and the creation of AI builder teams that are fully focused on outcome-driven delivery.”
The Macro Boom and Micro Gains
The momentum behind this shift is backed by staggering capital expenditure across the industry. Bhalla points out that “A McKinsey report notes that the rapid rise of AI is expected to drive investment of up to $7 trillion in global data centre and cloud infrastructure by 2030.” He adds that “In parallel, the four largest hyperscalers in the US are projected to invest close to $700 billion in AI infrastructure in 2026, buoyed by the surge in demand for cloud services.”
At the enterprise level, this infrastructure engine is translating into tangible operational advantages. “At Cognizant, we are starting to realize the benefits of agentic AI across parts of our business including Business Operations Automation, Service Desk operations, and software testing,” Bhalla shares. “When combined with intelligent software development and APOps workflows, these capabilities are delivering incremental productivity improvements beyond what classical automation alone can achieve today.”
The real transformation, however, occurs when the technology alters the end-user dynamic. The traditional IT service desk is being completely reimagined as a unified, conversational interface for both help and action.
“Rather than simply responding to queries, modern service desks are increasingly designed to resolve issues end-to-end through context-aware, multilingual interactions. At Cognizant, this shift is enabled through the Cognizant WorkNEXT suite, which brings together AI powered digital workplace capabilities to automate routine support and allow human teams to focus on complex exceptions and higher value work,” Bhalla explains. “Users benefit from a unified interface, faster outcomes, and language flexibility, while workflows are consolidated into a single intake with no handoffs, and productivity gains are driven by fewer tickets and a reduced cost to serve.”
The Rules Of Autonomous Governance
When decision-making shifts from humans to autonomous machines, traditional risk and compliance models break. “Autonomous agents require robust governance, as traditional models designed for human users alone are insufficient for agent-driven systems,” says Bhalla. To mitigate this, Cognizant’s approach is anchored in three core principles: identity-first governance, runtime enforcement, and outcome accountability.
Regarding Identity-First Governance, Bhalla emphasizes that “every agent is treated as a non-human identity that is formally registered, granted least-privilege access, continuously monitored and fully auditable. organisations need end-to-end traceability of actions and outcomes, because without it, autonomous systems can amplify risks rather than efficiency.”
The next layer is Runtime Enforcement, which “shifts governance from periodic audits to continuous enforcement. When an agent approves a transaction, triggers a workflow, or modifies infrastructure, guardrails, data permissions, and audit trails are enforced at execution.”
Finally, the system relies on Outcome Accountability as a key differentiator for Cognizant. “Our Autonomous IT Operations model embeds governance directly into the agentic delivery fabric, ensuring that compliance, risk controls, and policy enforcement are integral to how autonomous operations are designed, executed, and measured.”
The Composable Path Forward
As AI workloads surge, the cloud is entering a structural growth phase that rejects the traditional “lift-and-shift” playbook.
“Cloud is clearly entering a new phase of growth, driven by the rapid rise of AI workloads and the need for greater speed, scalability, and operational efficiency,” Bhalla advises. “This represents a structural shift, where enterprises must move beyond infrastructure-led spending to outcome-driven design. organisations should prioritise composable, agent-ready architectures rather than traditional lift-and-shift approaches. These modern designs enable faster optimization, better performance, and the ability to embed AI seamlessly across operations.”
Furthermore, a pragmatic multi-cloud strategy is vital to leverage the unique strengths of various hyperscalers while minimizing vendor lock-in. “This requires aligning architectural choices with workload characteristics, data gravity, cost optimization, and regulatory requirements,” Bhalla remarks.
Ultimately, the technology is only as good as the foundational skills and data assets supporting it. “Finally, talent and high-quality enterprise data are critical to success. organisations must prioritise upskilling engineers to effectively collaborate with AI systems and take greater ownership of end-to-end value delivery.”
The mandate for the modern C-suite is clear, and success hinges on a deep operational commitment. As Bhalla warns: “Real value emerges when AI becomes part of the operating fabric of the enterprise, not just an insight layer.”