American business executives are bullish on artificial intelligence (AI), and making their bets primarily on AI agents.
In fact, more than a quarter plan AI-related budget increases of at least 26%, according to PwC’s AI Agent Survey released Friday. An astounding 88% of companies plan to increase AI-related budgets in the next 12 months because of agentic AI.
The survey of more than 300 U.S. executives, taken between April 22-28, illustrates that in a choppy economy most enterprises are turning to AI agents to cut costs though few are using them to fundamentally rethink how to run their businesses.
Some 79% of senior executives are adopting AI agents into their companies, and nearly two-thirds (66%) report increased productivity. More than half (57%) have seen cost savings, faster decision-making (55%) and improved customer experiences (54%).
“Facing continued pressure to do more with less, many enterprises are turning to AI agents to boost productivity and reduce costs — and the early returns are promising,” Dan Priest, Chief AI Officer at PwC US, said in an email. “But the greatest potential of AI agents isn’t just efficiency; it’s reinvention. Too often, companies are deploying agents at the edges. The real leaders will go further — using AI agents to rethink processes, reshape operating models, and reimagine how work gets done.”
Still, adoption of AI agents hasn’t been entirely smooth, PwC researchers discovered in their survey. Less than half (45%) of companies are rethinking operating models and two-thirds said less than half of their employees interact with agents in everyday work.
“As with AI generally, early value often comes from internal use cases, but customer-facing cases are rising fast,” PwC said in its survey. “The danger here is in settling for too little. Companies that stop at pilot projects will soon find themselves outpaced by competitors willing to redesign how work gets done. Forty-six percent of respondents say they are concerned their company may be falling behind competitors in adopting AI agents.”
What is more, some senior business leaders are holding back AI agents over cybersecurity worries (34%) and costs associated with implementing the technology (also 34%) — even though a well-designed implementation can pay for itself and then some within months. The real challenges, PwC concluded, are rooted in organizational change such as lack of trust in AI agents (28%) and data issues (24%).
PwC’s survey is as much for internal use as it is intended for third parties. The consultancy, which unveiled two advanced AI tools for clients and a new research AI tool for its workforce on Friday, announced its own $1 billion, three-year investment in AI in April 2023.
Agentic work is currently the fastest growing area of PwC’s consulting services, it said, with more than half of its customer wins with Fortune 500 companies. It says 79% of its 75,000 U.S. employees use AI tools daily, and it has deployed more than 250 AI agents internally.
On Friday, PwC introduced a new business-performance engine using AI agents to help organizations drive enterprise-wide transformation through the combination of Large Language Models (LLMs), agentic AI and PwC’s industry models. The company also announced native support for the Model Context Protocol (MCP) in PwC’s agent OS that lets AI agents securely interact with tools and data through standardized, enterprise-ready connections.
