The trillion-dollar technology sector directly makes up an estimated 9% of the U.S. economy, indirectly even more as its innovations shape operational processes, products and services across industries and sectors. That technological diffusion is crucial to long-term economic growth.
But even as the pace of innovation quickens—cue agentic artificial intelligence—the process of spillover takes time. Many novel technologies are first used by the tech firms or organizations that invent them, or by a niche of early adopters. Other industries typically wait for commercial applications to see what the benefits are, what early proponents know and experience, and whether their business culture is open to change before climbing on board. Think of the internet’s origins in the 1960s as a way for Cold War-era government researchers to securely share information, and the later explosion in the 1990s when commercial browsers emerged—a historical arc shaped like the letter S.
Nowhere is that S-shaped trajectory more evident than with agentic AI, the fully autonomous offshoot of generative AI (gen AI) that sidesteps humans entirely to simulate and replicate their inputs, analyses and actions.
The rise of agentic AI
U.S. technology companies are far ahead of goods and services firms in knowing about agentic AI and feeling good about its potential to reshape business operations and consumer life. Awareness and knowledge from progressive exposure breed action, which is why tech firms populate the lower-left side of an S curve depicting growing adoption over time—akin to seedlings becoming a forest. Following the S-shaped trajectory implied in the cycle of innovation and “creative destruction” first described by influential economist Joseph Schumpeter in the 1940s, agentic AI is transitioning from rapidly evolving invention to the innovation of commercial products that spread—incrementally, and sometimes not yet at all—beyond the companies creating them.
These are just some of the findings in “Tech on Tech: How the Technology Sector Is Powering Agentic AI Adoption,” a PYMNTS Intelligence exclusive report. This edition examines how enterprises are viewing and reacting to the nascent rise of agentic AI. It draws on responses from 60 verified chief product officers and heads of product working at U.S. firms that generated at least $1 billion in revenues last year. The survey was conducted by phone from June 4, 2025, to June 17, 2025.
Tech Firms Are the Vanguard of Driving the Nascent Agentic AI Revolution
Full of the engineering talent, funding, investment budgets and long-term timelines needed to develop innovative software and its real-world applications, tech enterprises are far ahead of the curve in grasping agentic AI’s inner workings. Three in four are extremely familiar with the cutting-edge software, with the remaining one in four not far behind at very familiar.
The tech industry significantly outpaces the goods and service sectors in terms of its high-level knowledge. One in three goods firms are extremely familiar with agentic AI—less than half the level of tech firms—and the remaining two in three are very familiar. Services companies are coming up the curve but are still behind tech—38% are extremely familiar, while 62% are very familiar.
The gap reflects how reduced familiarity, a function of reduced exposure, stymies adoption, and highlights how companies “learn.” That process may become permanent. Bank of America wrote in a January research note that “we may be approaching an accelerating super cycle of innovation, in which increasingly powerful models produce increasingly powerful apps that in turn, drive additional model advances and app capabilities.” That suggests it’s possible that a “super cycle” of rapid improvements in agentic tools could lengthen the beginning of the S curve of adoption.
Some 42% of tech firms are exploring how they can bring agentic AI into their operations. Fewer than 4% of goods firms are. No services companies are. The technology has been hyped ever since OpenAI debuted in January an autonomous AI agent that can fill out online forms, make grocery lists and generate memes. Despite the buzz, even tech firms are treading relatively cautiously, suggesting that companies across all sectors are assessing, to varying degrees, their financial constraints, skill gaps, resistance to change and external factors including regulatory hurdles.
A Good Return on Investment in Gen AI Drives Trust in Agentic AI
Figure skaters want to nail a double axel before they try a triple. Similarly, companies that see a solid return on their investments in generative AI, agentic AI’s closely related but partially automated cousin, are the first in line to place their trust in agentic AI. Still, for tech firms, the nature of that ROI has shifted over the past two or so years.
In March 2024, just over one in two tech companies surveyed by PYMNTS Intelligence reported “very positive” return on investment (ROI) from using gen AI in their operations over the prior 12 months. By May 2025, only 17% did.
Nearly four in 10 tech firms reported a “somewhat positive” ROI over 12 months leading up to March 2024. Fourteen months later, that spiked to one in two.
Meanwhile, just 7.7% of tech firms saw “negligible” returns in the first survey period. That swung more positively by May 2025, when one in four did.
Overall, the survey data could suggest that tech firms went all-in on gen AI as hype around the technology first swelled, reaping bigger returns earlier in the process on earlier model iterations. However, as the rapidly evolving software continues to develop, firms may not see the same ROI they experienced in prior years. It’s possible that the returns timeline could in some ways mirror an S-shaped adoption timeline, requiring years for benefits to emerge.
Confidence in Agentic AI Varies According to What a Company Does
Agentic AI promises to do a lot of things for everybody without human intervention—write software, design new consumer products, ferret out better or cheaper overseas suppliers. But whether a company trusts the software to get the job done accurately and without bias hinges on what sector a company is in.
Goods-producing companies focus heavily on supply chain integration efficiency and sustainability, a reflection of their operational complexity and dependence on physical supply chains.
Service-sector firms emphasize cross-platform compatibility and the need for human oversight, reflecting their reliance on digital systems and their customer-facing operations.
Tech firms—already more familiar and confident with new forms of AI—focus on bias monitoring, evincing a more advanced AI literacy and heightened concern for reputation. Agentic AI claims for itself one superpower—replicating human thought and action without humans—but it has many applications and potential risks, leaving broad sectors of the economy still undecided about how to proceed.
Read More
PYMNTS Intelligence is the leading provider of information on the consumer trends driving innovation in consumer finance, digital payments and financial inclusion. To stay up to date, subscribe to our newsletters and read our in-depth reports.
Methodology
“Tech on Tech: How the Technology Sector Is Powering Agentic AI Adoption,” is based on 60 responses from verified chief product officers and heads of product working at U.S. firms that made at least $1 billion in revenues last year. The survey was conducted by phone from June 4, 2025, to June 17, 2025.
