From SaaS to agents: Why AWS believes Indian software firms are ready for another transition

For more than a decade, Indian software companies built businesses around a relatively straightforward idea: deliver software over the cloud. That model helped create some of the country’s most successful enterprise technology companies. Firms such as Freshworks, CleverTap, Amagi and C.E. Info Systems now serve customers across global markets, including large multinational enterprises.

AWS believes another shift is underway. “If you go back about 10 or 12 years, SaaS as a concept came up. People built software applications on top of cloud technologies. Today, they are building agents on top of cloud technologies,” says Praveen Sridhar, head of partner and ISV business at AWS India and South Asia.

In AWS’s view, agentic AI is not replacing software in the way many industry observers suggest, and is instead becoming another layer in the technology stack. Sridhar describes the progression as a series of additions rather than disruptions. Infrastructure moved to the cloud. Foundation models emerged on top of cloud infrastructure. Agents are now being built on top of those models. “Earlier there was SaaS. Now on top of SaaS, people are bringing in models and agents. It is simply making the service they were already providing to customers more intelligent,” he says.

That distinction matters because much of the recent debate around enterprise AI has centred on whether traditional software companies risk becoming obsolete. AWS sees the situation differently. Many of the companies now building AI agents are the same firms that spent the past decade building software products.

India's Agentic AI transformation 

AWS’s Agent Marketplace, where businesses can list and distribute agent-based applications, now hosts more than 3,000 AI agent solutions globally. According to the company, it has become the fastest-growing category in Marketplace history. “The next evolution of software is through the format of an agent,” Sridhar says.

For Indian software companies, the transition comes at a familiar moment. The country’s SaaS sector was built on a combination of engineering talent, cost efficiency and the ability to serve customers across geographies. AWS expects many of those advantages to carry over into the AI era. “Indian innovators and Indian companies are extremely innovative and frugal in nature. Just as Indian founders that built solutions in the SaaS world, they are now going to build agents in the agent world.”

AWS is positioning Bedrock, its managed service for foundation models, as a key platform for that shift. Companies including Freshworks and C.E. Info Systems are already working with the service, according to Sridhar. AWS is also promoting AgentCore, a platform aimed at helping software vendors build and deploy AI agents.

AI talent and data readiness shortage as challenges

The opportunity is large, but so are the execution challenges. Sridhar points to a finding from McKinsey’s State of AI report: while 88% of organisations are using AI in some form, only 39% report a measurable impact on earnings.

The issue, he argues, is rarely the technology itself. Many organisations still struggle with data readiness. Others face shortages of AI-skilled talent. Integrating new AI systems into existing enterprise software remains difficult. Then comes the challenge of scaling a successful pilot into a production environment. “It is one thing to run a small pilot. It is another thing entirely to deploy it in production,” he says.

Those challenges are becoming increasingly important as companies attempt to justify rising AI investments. Amazon has committed $12.7 billion towards cloud infrastructure in India by 2030 as part of broader investments across the country. Yet questions around returns on AI spending continue to surface across boardrooms.

Sridhar’s argument is that companies often start in the wrong place. “You should not take up an AI project simply because everyone else is taking up an AI project,” he says. “First, ask whether you have the foundations in place.” That means identifying a business problem before selecting a technology.

For Indian software firms, the transition from SaaS to agents may ultimately depend less on the underlying AI models and more on a familiar challenge: solving real customer problems.