Google Cloud study says 52% of firms use AI agents

More than half of large organisations have deployed artificial intelligence agents, with early adopters reporting stronger financial returns and broader operational impact.

The research, based on a survey of 3,466 senior executives across 24 countries, shows that 52% of organisations are using AI agents, while 39% have deployed more than ten across their operations.

AI agents - systems that can plan and execute tasks with limited human input - are shifting from experimental tools to core business systems. Adoption spans industries and regions, with use cases across customer service, marketing, cybersecurity and software development.

A group identified as "agentic AI early adopters" represents 13% of respondents. These organisations allocate at least half of future AI budgets to agents and have integrated them deeply into operations. Among this group, 88% report a return on investment from generative AI in at least one use case, compared with 74% overall.

"This year's research shows we're entering the next chapter of the AI wave. The conversation has moved from 'if' to 'how fast,' and the new differentiator is agentic AI," said Oliver Parker, Vice President, Global Generative AI Go-To-Market, Google Cloud. "Early adopters of agents are not just automating tasks; they are also redesigning core business processes. By championing AI as a core engine for competitive growth and thus securing dedicated budgets, they are providing a clear roadmap for any organization looking to scale, solve complex challenges, and achieve more consistent ROI."

AI agents are used across multiple functions, most commonly in customer service, marketing, security operations and technical support.

Use varies by industry. Financial services firms focus on fraud detection, while retailers prioritise quality control. Telecommunications companies emphasise network configuration. Regional differences are also evident. European organisations favour AI-driven technical support, while companies in Asia-Pacific focus more on customer service.

"We're seeing organizations around the world use agentic AI to tackle complex industry-specific tasks – from fraud detection in financial services to quality control in retail," said Carrie Tharp, Vice President, Head of Strategic Industries and Solutions, Google Cloud. "This isn't just about efficiency; it's about embedding intelligence directly into the business."

Financial returns from generative AI remain consistent. Around 74% of executives report achieving a return on investment within the first year.

More than half say AI has contributed to business growth. Among those reporting gains, 71% cite increased revenue, with 53% estimating growth between 6% and 10%.

Executives also report improvements in productivity, customer experience and time to market. A majority cite productivity gains, while many report better customer outcomes and measurable business growth.

Investment in AI is increasing as costs decline. About 77% of executives report higher spending on generative AI, and nearly half are reallocating budgets from non-AI projects.

Implementation timelines are shortening. More than half of organisations now move from concept to production within three to six months.

These patterns point to a shift from experimentation to scaling, with AI becoming embedded in operational workflows and aligned more closely with business objectives.

As adoption grows, concerns around data protection and system integration are becoming more prominent. More than one-third of executives rank data privacy and security among their top considerations when selecting large language model providers.

Integration with existing systems and cost remain key factors. Organisations are focusing on governance and infrastructure before adopting more advanced capabilities.

"2024 proved that generative AI works; 2025 is all about compounding that success," said Parker. "The biggest hurdles for most organizations are rooted in foundational data security and systems integration. The solution is to adopt a modern data strategy with strong governance from the start."