GenAI disrupts software ADM market, 10-15% of IT services revenue at risk

While revenue growth from AI is still marginal, the larger question is how the industry can reinvent itself when one of its largest revenue pools becomes less labour-intensive

As the debate continues over the revenue-generating potential of generative AI (GenAI), one thing is becoming increasingly clear: its impact on the Software Development Life Cycle (SDLC) — and by extension, Application Development and Maintenance (ADM) — is already being felt in measurable ways.

 

According to a report from Motilal Oswal Financial Services, ADM, which makes up an estimated 35–45 per cent of the IT services industry’s revenue, is emerging as ground zero for GenAI’s most immediate and tangible productivity gains.

 

For IT services firms, this translates to a structural challenge. "Our research suggests an approximately 40 per cent productivity gain from enterprise-wide implementation of GenAI Copilot, putting about 10-15 per cent of IT services revenues at risk,” wrote the report's authors Abhishek Pathak, Keval Bhagat, and Tushar Dhonde.

 

Another report from Kotak Institutional Equites also added that adoption of AI has increased in select use cases. The report also noted that efficiencies from AI adoption in use cases such as software engineering can be significant.

 

“Copilot with GitHub and Claude are commonly used tools for coding. Use cases are pretty high in application services (both development and maintenance), content generation and BPO services. Clients are not resisting usage of generative AI tools due to concerns around the technology; instead, they are encouraging vendors to adopt generative AI in such use cases,” the Kotak report pointed out.

 

Some of the work within SDLC that is being immediately impacted includes low-level coding or routine feature work, code review and testing, debugging & incident response, and security fixes. In a recent story by Business Standard on impact of AI on testing, it was found that as more and more code is being written by AI, it has raised questions about the future of traditional testing engineers, establishing the need for them to reskill quickly to stay relevant.

 

Motilal Oswal’s analysis noted that GenAI tools like Copilot drive around 55 per cent efficiency in repetitive coding tasks, translating to 11 per cent of total ADM hours saved. Similarly, automated suggestions and AI-generated test cases reduce effort by about 40 per cent, resulting in overall time savings of 8 per cent.

 

The findings were discussed in a recent industry session featuring Saurabh Gupta from HFS Research, who emphasised that the days of relying on anecdotal GenAI success stories are over. “Large enterprises are encountering the law of diminishing returns. The limits of offshoring have been reached, and jumping into a new "S-curve" of value creation is increasing,” he said.

 

In the near term, companies that fail to deploy GenAI internally at scale risk being undercut by more efficient rivals. But in the long run, the bigger question is how the industry can reinvent itself when one of its largest revenue pools becomes significantly less labour-intensive.

 

Analysts also highlighted that the current shift towards AI is replicating the shift the industry went through from legacy to digital. The industry is facing productivity and pricing pressure in the ADM and core IT services, but the revenue uplift from GenAI is yet to happen.