From 4 weeks to 5 days: How Infosys’ AI agents are boosting enterprise operations

Infosys

Early deployments of Infosys’ agentic artificial intelligence (AI) systems are delivering measurable gains for enterprise clients, cutting audit turnaround times from four weeks to five days, and automating complex workflows across industries.

The company’s AI agents, autonomous software capable of decision-making, are built using open architecture, supported by Infosys’ Topaz platform and Google Cloud’s Vertex AI. They are being applied to areas such as finance operations, customer service, and supply chain management.

“Agentic AI is like the ChatGPT moment for enterprises,” Balakrishna DR, executive vice president and global services head, AI and industry verticals at Infosys, told Moneycontrol at the launch of firm’s research report Responsible Enterprise AI in the Agentic Era Radar.

“You see a clear business case for implementing it, improving processes, bringing in efficiency, reducing costs, and enhancing customer experience,” he said.

The report from the Infosys Knowledge Institute finds that while 78 percent of business leaders believe responsible AI (RAI) drives growth, only 2 percent of enterprises meet benchmark maturity standards.

Nearly all respondents (95 percent) reported AI-related incidents in the past two years, with 40 percent of those severe enough to threaten business continuity.

Audit work in 5 days

Citing some examples of agentic AI, Balakrishna said a leading professional services firm tasked Infosys to revamp its audit workflows. What once took four weeks was done in just five days, thanks to enterprise AI agents that automated routine tasks.

This isn’t a one-off story.

From invoice automation at Americana Restaurants, which serves 2,600 global locations, to retailer onboarding for a major CPG firm, Infosys’ AI agents are embedded into the flow of enterprise work.

In each case, the agents handle real-world complexity: exceptions, multilingual logic, dynamic decision-making and cross-functional coordination.

In the CPG use case, for instance, onboarding time was cut to half and manual effort dropped by 50 percent.

Human in the loop

While agentic AI offers automation benefits, it also introduces new risks. Infosys said it has taken early steps to embed RAI into its development and deployment processes.

“We created a responsible AI office well before scaling up agentic systems,” Balakrishna added. “We built guardrails into the platforms, policy restrictions, red-teaming, adversarial testing and human-in-the-loop oversight.”

As agentic AI adoption accelerates, Infosys is positioning itself as both an innovator and a systems integrator. Through its Foundry platform, the company hopes to enable enterprises to customise and operationalise intelligent agents rapidly, without compromising oversight or compliance.

“We’re not building for a one-off,” he said. “These agents must scale across industries, be secure by design, and adapt to different enterprise environments.”