TL;DR

New research from Gallup reveals AI adoption may be flatlining after initial enthusiasm, with the total number of AI users remaining unchanged and 49% of US workers reporting they never use AI at work. Separate research from Hitachi Vantara found 95% of organisations are seeing zero return on generative AI investments due to infrastructure gaps.

Adoption Plateau Emerges

Gallup’s Q4 2025 research found daily AI use increased just two percentage points and frequent use rose only three percentage points quarter-over-quarter. Despite considerable corporate investment and media attention, half of the US workforce reports never using AI in their roles.

Clear sector biases are evident, with technology, finance, and professional services leading adoption, alongside education. The biggest growth came from ‘remote-capable’ desk-based roles, more than doubling from 28% in 2023 to 66% in 2025. Non-remote-capable roles lag approximately two years behind at 32%.

The Leadership-Worker Gap

There’s a significant divide in AI usage by seniority. Leaders (69%) and managers (55%) are considerably more likely to use AI compared with individual contributors (40%). This gap suggests AI tools may be filtering down slowly through organisations, or that frontline workers see less relevance in current applications.

The ROI Challenge

Hitachi Vantara research reveals a stark deployment-versus-usage divide. While 98% of companies are using, piloting, or exploring AI, a striking 95% are getting zero return on their generative AI investments due to infrastructure gaps.

Only 42% of organisations consider themselves data-mature, despite high-quality data being identified as the top driver for AI success. ‘AI succeeds when the data behind it is trusted, well-governed and resilient,’ noted Hitachi Vantara CEO Sheila Rohra.

Looking Forward

The findings suggest existing AI users will increase their usage frequency before broader workforce adoption occurs. Clear use cases, rather than speculative investments, will drive meaningful growth. Trust, security, and governance must become priorities to address concerns about data breaches, unreliable outputs, and regulatory risks that continue to hamper enterprise AI adoption.