Infosys Research Finds Growing Gap Between AI Investment and Measurable ROI

Infosys research finds that companies are struggling to measure AI ROI, with efficiency, cost savings and faster time-to-market emerging as key benefits.
Infosys Research Finds Growing Gap Between AI Investment and Measurable ROI

Bengaluru, August 25, 2026: Infosys has released a new research report highlighting a growing challenge for businesses investing heavily in artificial intelligence: turning AI spending into measurable enterprise value.

The report, titled “The AI ROI Gap: Turning Ambition Into Enterprise Value,” is based on a survey of more than 1,000 senior executives at large US companies.

While businesses continue to view AI as an important driver of future growth, the research suggests that many are finding it difficult to demonstrate returns from their investments.

Key Highlights

  • 80% of executives believe AI is a critical driver of new revenue opportunities.

  • One in four executives say AI ROI has fallen below expectations.

  • 75% of respondents report that AI is generating a net positive value for their organisation.

  • Two-thirds say their organisations struggle to measure AI ROI effectively.

  • Nearly half do not have a centralised KPI framework for measuring AI outcomes.

  • Nearly three-quarters say fewer than 25% of AI pilots have successfully scaled across the enterprise while delivering their intended ROI.

  • Executives identified unclear business cases and poorly defined ROI targets as major barriers to scaling AI projects.

 

Follow PSU Connect:WhatsApp

 

AI’s Biggest Gains May Be Coming From Efficiency

According to the research, companies may be looking at AI returns too narrowly by focusing primarily on revenue growth.

The strongest measurable benefits are currently appearing in areas such as faster time-to-market, operational efficiency and cost reduction.

This suggests that AI is increasingly functioning as a tool for improving internal business performance rather than immediately producing direct top-line revenue.

 

Measuring AI Value Remains a Challenge

Another major issue highlighted by the study is measurement.

Although companies are investing heavily in AI, many lack consistent systems for determining whether those investments are delivering the expected results. The report notes that only around one-quarter of executives formally track speed-to-market improvements, despite identifying it as one of AI’s strongest areas of impact.

As a result, some organisations could be undercounting the value already being created by AI.

Follow PSU Connect:LinkedIn

 

Scaling AI Pilots Is Still Difficult

Moving from an AI experiment to an enterprise-wide deployment remains another major hurdle.

The research found that a large majority of respondents said fewer than 25% of AI pilots have successfully reached enterprise scale while delivering their intended ROI.

Infosys says companies need to establish the business outcome and ROI expectations before selecting and deploying AI technology.

Governance and People Matter Too

The research also points to growing concerns around employee overreliance on AI, unauthorised use of AI tools and security risks.

As adoption increases, companies may therefore need to focus not only on technology but also on employee training, governance and human oversight.

What Companies Can Do

Infosys recommends that organisations looking to close the AI ROI gap should:

  1. Define clear business outcomes before deploying AI.

  2. Measure value beyond direct revenue generation.

  3. Build stronger KPI and ROI measurement frameworks.

  4. Improve governance around AI adoption.

  5. Maintain appropriate human oversight and employee training.

  6. Focus on scaling AI projects that demonstrate clear business value.

Bottom line: Infosys’ research suggests that the AI challenge is shifting from “How much should companies invest?” to “How can they prove the investment is creating value?” Companies that combine ambitious AI adoption with disciplined measurement, governance and execution could be better positioned to turn AI experimentation into sustainable enterprise value.

Note*: This article is for informational purposes only. PSU Connect is not responsible for any actions taken based on this content.Terms & Conditions