Enterprises are under mounting pressure from boardrooms to factory floors to move beyond AI pilots and proofs of concept. The conversation has shifted from potential to performance, from experimentation to execution. AI is now expected to deliver measurable business outcomes, fast. At the centre of this shift is Agentic AI, a new class of systems that go beyond support roles to act autonomously, make decisions and generate value without constant human input.

In HCLTech’s latest report, The Blueprint to an AI-led Operating Model, one thing is clear: Agentic AI isn’t an experiment or a future vision; it’s the next operating standard. These systems are evolving from assistants to autonomous actors, capable of adapting in real time and driving outcomes across the enterprise. This transition is already reshaping how organizations approach productivity, decision-making and growth, while sharpening the focus on return on investment.

But what does that transformation look like in practice? How are enterprises adopting this intelligence and what barriers must they overcome to realize its full value?

In a recent Trends and Insights discussion, Piyush Saxena, SVP and Global Head of the Google Cloud Business Unit at HCLTech, offered a direct, real-world perspective, grounded in deployment experience, customer outcomes and product innovation.

“It all began with robotic process automation, where we streamlined repetitive tasks using structured data, like report generation, invoice processing and data entry. Then came GenAI, which expanded our capabilities by working with both structured and unstructured data, identifying patterns and creating original content. Today, we’ve stepped into the era of Agentic AI, systems that execute tasks independently and make decisions in dynamic, real-world environments,” said Saxena.

An agent behaves like a human counterpart, adjusting to changes, acting in real time and learning through incomplete data sets. Think of an autonomous vehicle that must respond instantly to a pedestrian stepping into the street. The agent doesn’t wait for a command; it acts, safely and smartly. That’s the kind of autonomy enterprises now seek within their systems.

From ideas to impact: Where enterprise customers stand today

Saxena believes the agentic shift is already well underway. “In the Google Cloud space, nearly 100% of our customers are engaged in some shape or form of Agentic AI,” he said. “Some are still identifying use cases, some are running agents in controlled environments and a few are already in production.”

This variation reflects a growing maturity curve but also highlights shared challenges.

“Identifying the right use case is the biggest hurdle. It sounds simple, but it’s not. Once that’s done, organizations need to go through multiple layers of security, compliance and business approvals. Then comes the complexity of securing data environments and delivering ROI,” explained Saxena.

The interest is real, but so is the complexity. Many enterprises are realizing that Agentic AI is not a plug-and-play solution. It requires foundational work like clean data, clarity on business goals and cultural readiness to achieve machine autonomy. This is why organizations that succeed often take a phased approach, starting with one high-impact use case that can demonstrate tangible value and gain executive sponsorship before expanding to broader transformation programs.

The question of return on investment sits front and center. “Business leaders are asking — what do I get out of this? How does this drive revenue, or optimize cost,” he said.

Making ROI real: Agentic AI in action

HCLTech isn’t just advising customers on AI adoption, it’s helping them deploy at scale. The company has published 50 industry-aligned agents on the Google Cloud Marketplace, with another 200 available for enterprise-specific deployments.

One standout case comes from the manufacturing sector.

“We’re working with a North America-based paper manufacturer,” shared Saxena. “We built a conversational AI agent that lets engineers interact with machine setup systems using natural language. It helps them troubleshoot and resolve faults faster. The impact? Thirty percent faster machine setup time and an estimated $170,000 per plant per year in ROI.”

The tech stack behind this includes Vertex AI, Agentspace and Gemini 2.5, integrated into HCLTech’s Insight solution for predictive defect detection.

Saxena broke down the ROI formula in simple terms: “We start by assessing the revenue impact — how much more value the organization can generate and how quickly. Then we factor in three cost components: building the agent, running it on Google Cloud and supporting it over time. The difference between the value created and these costs is your ROI.”

That level of transparency in measurement is key to building trust with stakeholders and sustaining momentum for AI investments.

A winning alliance: HCLTech and Google Cloud

The innovations described above are grounded in a long-standing partnership between HCLTech and Google Cloud. This relationship has evolved since 2019, when HCLTech became one of the first partners to launch a dedicated Google Ecosystem Business Unit. Since then, HCLTech has won seven Partner of the Year awards, including the Global Talent Partner of the Year for two consecutive years.

“This is a very strategic partnership, we work closely with Google Cloud product teams on emerging technologies and we’re focused on building industry solutions around them,” said Saxena.

That focus on domain-specific outcomes has been a major differentiator. “Our goal isn’t to sell technology, it’s to solve industry problems. For instance, HCLTech Insight helps manufacturers detect and correct defects. HCLTech Netsight is designed for telcos to optimize their networks. We also have tailored solutions for financial services and life sciences. That’s the first differentiator.”

The second differentiator is depth. HCLTech has built a dedicated Center of Excellence solely focused on Google Cloud technologies. This team not only gains early access to product updates and capabilities but also plays a key role in shaping deployment strategies, often staying ahead of the curve in bringing new solutions to life.

Complementing this is the scale of talent. With over 20,000 Google Cloud-certified engineers across geographies and industries, HCLTech offers clients access to specialized expertise at every stage of their transformation journey, from advisory to implementation to ongoing support. Together, this depth and scale give the organization a distinctive edge, not just in delivering technically sound agentic systems, but in building trust with enterprise stakeholders who need reliable, scalable solutions aligned to real industry outcomes.

What to expect at Google Cloud Next London

HCLTech’s partnership with Google Cloud will be on full display at the upcoming Google Cloud Summit in London, where Saxena and his team will showcase live demos of agentic use cases across industries.

“We’re pretty excited about the summit and we’ll be showcasing a range of AI agents; supply chain management agents, fraud detection agents, trade surveillance for financial services and adverse event intake for life sciences,” said Saxena.

Attendees can also engage directly with HCLTech’s domain and technology experts. “We’re not just presenting demos, we’re opening the floor to discussions. Customers will have the chance to talk about their challenges and learn from our deployment experience,” he added.

A special area of focus will be the Agent-to-Agent (A2A) protocol, a new frontier in distributed autonomy where HCLTech is a Google Cloud launch partner. “We’re actively contributing to A2A and the summit is a great place for customers to see what that future might look like,” said Saxena.

From possibility to production: The agentic enterprise is here

As the conversation wrapped, Saxena returned to a theme that runs through HCLTech’s AI strategy, outcome-centric innovation.

“Agentic AI is not about trends, it’s about delivering impact. It’s about systems that not only think but act, not only respond but adapt. And when done right, they don’t just support the business, they become part of how the business runs,” he mentioned.

For enterprises seeking to redefine how work gets done, where insights are actioned and how value is created, Agentic AI is the blueprint for what’s next.