Study Finds Enterprises Rethink Work to Get More From AI

Enterprises have spent millions of dollars acquiring artificial intelligence models, copilots and capabilities. The next dollar thrown at a smarter model may be worth less than the next dollar spent fixing the data, permissions, workflows and organizational plumbing preventing existing AI from doing real work.

That is the bigger implication of the PYMNTS Intelligence report “AI at Work: Why Deeper Enterprise Use Produces Stronger Returns,” which showed in September a disconnect between how much AI companies use and how much value they extract from it.

Companies with AI embedded in one or two functions use it across an average of 41 of the 75 tasks measured, the report found. Companies with AI embedded in three or more functions use it across 40. Virtually no difference. However, 55% of the first group reported generating returns from AI investments. Among the deeper adopters, that figure jumped to 93%.

AI could be approaching a familiar technology inflection point. The bottleneck has moved. The internet required data centers and logistics networks. Cloud computing required application migration and new security architectures. Smartphones created downstream markets in payments, identity and mobile software.

What companies don’t necessarily have is an enterprise designed to let capable AI machines operate inside it. That becomes more consequential as AI moves from answering questions to performing work.

An AI assistant drafting a memo can sit on top of a messy organization. An agent approving an invoice, changing a customer account, initiating a procurement workflow or recommending where millions of dollars should move cannot.

Those sound like technology requirements. They’re actually operating model requirements.

That helps explain another seemingly strange finding in the PYMNTS Intelligence data. Companies with three or more embedded AI functions reported an average of 5.6 barriers to adoption, compared with just three among companies with nothing embedded. The companies furthest along with AI aren’t discovering fewer problems. They’re discovering more of the company.

Production AI exposes organizational debt that pilots can ignore, such as incompatible databases, fuzzy ownership, inconsistent policies, disconnected workflows and approval structures built for humans moving information manually from one system to another. As a result, the next enterprise AI spending boom may technically be counted as spending on cybersecurity, identity, cloud infrastructure, data management, consulting, integration or workflow software.

Economically, however, much of it could be AI-enablement spending.

If access to powerful models continues commoditizing, possessing AI itself won’t provide much competitive advantage. Competitors can buy access to roughly the same intelligence.

What they cannot buy overnight is the organizational architecture required to exploit it. That turns clean data, interoperable systems, clear decision rights and mature governance from corporate housekeeping into productive assets.

The question is no longer which company has the smartest model. Everyone can rent one. The question is which company can actually put it to work.

For all PYMNTS AI coverage, subscribe to the daily AI newsletter.

At PYMNTS Intelligence, we work with businesses to uncover insights that fuel intelligent, data-driven discussions on changing customer expectations, a more connected economy and the strategic shifts necessary to achieve outcomes. With rigorous research methodologies and unwavering commitment to objective quality, we offer trusted data to grow your business. As our partner, you’ll have access to our diverse team of PhDs, researchers, data analysts, number crunchers, subject matter veterans and editorial experts.