Why enterprises still struggle to scale AI beyond pilot projects

Why enterprises still struggle to scale AI beyond pilot projects

A Forrester study, commissioned by FPT Corporation, found that 51% of enterprises allocate at least 5% of IT budgets to AI, but only 26% are advanced in operationalizing it.

More than half of enterprises are allocating at least 5% of their IT budgets to artificial intelligence, while 26% consider themselves advanced in putting AI into operation, according to a new study commissioned by FPT Corporation.

The research, conducted by Forrester Consulting, surveyed 397 business and technology decision-makers and included interviews with senior executives across North America, Europe, Asia Pacific, and Japan. Respondents represented industries including automotive, financial services, healthcare, manufacturing, energy, and sports.

Among the organizations surveyed, 51% allocate at least 5% of their IT budgets to AI, while 26% consider themselves advanced in operationalizing it. AI is primarily being used for automation, cost reduction, and other operational purposes, while 34% of respondents said their organizations were pursuing an AI-first operating model.

Integration complexity was identified as a challenge by 41% of respondents, while 38% cited data silos. Another 39% of enterprises reported meaningful progress in aligning AI strategy, governance, and operating models.

The study found that 35% of respondents do not collect quantified metrics for AI initiatives, while 10% do not measure AI outcomes in any form.

According to the study, the lack of quantified metrics makes it harder to determine which AI initiatives should be expanded or deprioritized.

When selecting external AI partners, 48% of respondents prioritized the ability to engineer, deploy, and operate AI systems throughout their lifecycle, while the same proportion cited governance and security capabilities.

Integration with existing systems was identified as a priority by 47% of respondents. Full-lifecycle AI capabilities were prioritized by 59% of respondents in North America and 54% in Europe, the Middle East, and Africa.

Organizations in Asia Pacific and Japan placed greater emphasis on strategic and execution support across the deployment process, according to the study.

Pham Minh Tuan, executive vice president of FPT Corporation and CEO of FPT Software, said organizations need to address AI strategy, integration, governance, and operations together rather than as separate workstreams.

"Organizations can no longer move forward in silos—they need partners who can bridge strategy, integration, governance, and operations," he said.

FPT commissioned the study as part of its enterprise AI business. The company operates the FleziPT AI platform and AI facilities in Vietnam and Japan.

FPT claims its AI-supported development processes have reduced development time by up to 60%, cut rework by more than 50%, and increased developer productivity by 30%. FPT did not provide methodological details for those figures in the announcement.

The company has also introduced FPT CASAN, a methodology intended to assess organizations at five levels of AI adoption: Curious, Augmented, Standard, Automatic, and Native.

According to FPT, the framework is designed to assess AI readiness, governance, and deployment across business functions. The company said it is intended to help organizations move from individual AI experiments toward broader operational deployment.