2025 has become the year of agentic AI, marking the latest advance in AI technology after the still-evolving generative applications that preceded it. This particular development not only is taking hold, but it’s reaching scale at a rapid pace. More than half (52%) of 3,466 senior corporate leaders at companies that use generative AI, surveyed by Google Cloud and National Research Group, said they’ve also begun to leverage AI agents.
Google Cloud’s survey report defined AI agents as specialized large language models that independently plan, reason and perform tasks and can interact with other AI agents as needed. These agents, which have specific roles, context and objectives, can be pre-built or built in-house.
About 4 in 10 (39%) of the surveyed executives said their organization has launched more than 10 AI agents. That rises to 82% among early adopters, defined as those who have dedicated at least 50% of their future AI budgets to agents.
Use cases range from simple chatbots to complex, multi-agent systems that can independently manage complex workflows with limited human guidance.
According to the survey report, usage of agentic AI varies by region, from a 64% adoption rate in the Japan/Pacific region to 46% in North America. How the technology is used also varies, with tech support a key priority in Europe, customer service in Japan/Pacific and marketing in Latin America.
Worldwide, among organizational functions, customer service leads in use cases, with 49% of the survey participants saying their companies use AI agents in that area, closely followed by marketing, cybersecurity and tech support.
There is little differentiation by industry in the volume of usage, with most industries’ adoption rates sitting between 51% and 56%.
Interestingly, adoption rates also are fairly homogenous among different-sized companies: 49% among those with 100 to 499 full-time employees, 56% for those with 500 to 999 FTEs and 52% for larger employers.
Many companies are using agentic AI to help scale existing AI programs. Among early adopters, 88% have already experienced ROI on at least one generative AI use case, versus 74% across all organizations.
“Leaders in these organizations champion AI in production, sponsor the deployment of multiple agents across the business, and secure dedicated budgets for growth, translating into more consistent ROI,” Google Cloud wrote.
According to the report, early adopters of agentic AI are more likely to report significant value from generative AI across key areas, including customer experience, business growth, security and marketing.
