New SAP Study Questions Whether CX Businesses May Be Overpromising On AI

It seems retailers may have been overestimating how prepared they are for agentic commerce.

When a new SAP survey asked decision-makers in customer experience (CX), nine out of 10 said they are “at least somewhat prepared” to use AI agents to serve their customers.

This confidence, however, does not hold up after further questioning: Only 6 percent of U.S. and 3 percent of German CX decision-makers can say all the relevant systems to support end-to-end AI are ready.

It’s one of the main findings in the enterprise software company’s latest white paper. Dubbed the Autonomy Gap, it refers to the distance between what AI can do, what consumers are willing to delegate and what businesses can reliably deliver.

It’s not just businesses that struggle to keep up. SAP said improvements in AI capabilities are moving too fast for consumers to trust the technology. In a hypothetical situation where AI can shop on their behalf, 87 percent of U.S. and 83 percent of German consumers say they would still want to review and approve the final purchase.

“Customers may still make the final purchase decision, but AI is increasingly influencing which businesses they consider. That gives brands less control over how they’re evaluated, even as they remain responsible for delivering what customers choose,” SAP said.

It’s a reality that brands now have to contend with. In order to reach the consumer, they must first reach the AI, and that means making their content readable by AI systems. SAP said brands should make their products and services easier to compare “without becoming easier to replace.”

The survey also looked into what these transactions look like in business-to-business (B2B) arrangements. Within the same survey pool, SAP analyzed 2,923 AI buyers from the consumer sample and 749 B2B buyers from the CX practitioner and decision-maker sample.

Whereas consumers may be slow to trust, B2B buyers are quick to adopt. So much so that nearly all of them in the study said they use AI for business purchasing or vendor selection, with about a third describing the experience as very helpful.

“That’s meaningful early progress given the complexity of B2B transactions. For companies, it’s also a strong signal to prepare for AI playing a larger role in how customers buy,” SAP said.

The report also found a mismatch between what consumers and businesses prioritize when using AI in retail.

For example, most consumers in both markets say using AI for price comparisons and discounts is their number one priority. Businesses, on the other hand, rank this seventh. Meanwhile, the top AI priority for businesses is producing personalized recommendations, a function that ranks sixth for consumers.

“AI makes an old commercial tension much more visible: Businesses want growth and healthy margins. Customers want the right product or service at a price they believe represents fair value,” SAP said.

Nevertheless, SAP said enterprises have a responsibility to follow through on whatever their AI promises consumers.

“A brand can make a compelling promise,” SAP said. “When AI makes that promise on its behalf, the enterprise has to keep it.”