News

Jay Roxe is Chief Marketing Officer at Inriver.

getty​Earlier this year, my company fielded a benchmark study of 405 senior technology, data and marketing executives across industrial manufacturing and wholesale distribution. We asked two kinds of questions. First, how capable do you believe your product data operations are? Second, what do those operations actually deliver?

One hundred seventeen of those executives told us their organizations were fully prepared for agentic commerce, the coming era in which AI agents autonomously discover, evaluate and transact on behalf of buyers. When we tested those claims with operational measures of product data maturity, none of them cleared the bar. Zero out of 117.

I’ve been in marketing technology long enough to expect some gap between what executives claim in a survey and what their operations can back up. I did not expect the gap to be total. And the more we dug into the data, the more it looked like the gap itself was the story, because it doesn’t show up where you’d think.​

Confidence And Capability Are Moving In Opposite DirectionsThe self-assessments in our study were strong. On seven-point scales, respondents rated their core product data capabilities between 5.8 and 6.0, which are not the scores of organizations in distress. The operational answers told a different story. Only a third had more than 85% of their active SKUs publish-ready at any given moment. Nearly half deal with integration failures between systems monthly or more often. Fewer than 14% have built end-to-end automation with an audit trail.

You cannot square a near-top self-rating with integration failures running monthly for half the sample, and the consequences of that mismatch are not academic. When leadership allocates budget based on a self-assessment that operational reality doesn’t support, the money goes to the wrong problems, the real gaps stay unfunded, and nobody notices because nobody is measuring the thing that would reveal them.

What worried me most, though, was the direction of the relationship. Organizations with deployed product information infrastructure were meaningfully more likely to reach the upper operational tiers of our index. Yet, organizations without that infrastructure rated their AI readiness higher on every dimension we measured. The executives closest to the problem scored themselves modestly. The ones furthest from it scored themselves highest.

We saw the same inversion across industries. The vertical with the most rigorous measurement discipline in our sample, electronics, claimed AI readiness at the lowest rate of any industry while posting the least downstream friction, from channel disputes to launch delays. The industry claiming readiness at nearly double the field rate had the lowest automation levels in the study and reported the worst downstream pain.

Experience with the problem produces humility. Distance from it produces confidence.

Why Agents Change The StakesOverconfidence about data operations has been survivable for 20 years because humans absorbed the gap. A sales rep works around a sync delay. A channel manager spots the outdated spec sheet before the customer does. Someone calls support, and someone else fixes it.

AI agents don’t work around anything. An agent querying your product catalog receives whatever your integration returns at that moment and acts on it, and if what it receives is incomplete or stale, it may move on to a competitor whose data answers cleanly. The manual heroics that make a fragile system look functional from the inside are invisible to an agent, and worthless to it.

It’s worth noting what the small number of top-tier organizations in our study had in common. None of them set out to become “AI-ready.” They set out to fix integration failures, automate product launches, and measure what their content actually returned. Readiness showed up as a by-product of that discipline.

3 Questions To Ask This QuarterSince confidence can’t be trusted as a signal, I’d push any leadership team toward operational answers instead.

• What percentage of active SKUs are publish-ready right now, today, not averaged over the quarter? If the answer is, “We don’t know,” that is itself the answer, and it puts you in the majority of our respondents.

• When did your most recent integration failure occur? If the answer is, “We don’t track that,” the failures are still happening. They’re just being absorbed, invisibly, by the people downstream of them.

• Which of your AI claims would survive a “show me production” meeting? Whatever doesn’t survive that meeting is your actual road map.

The Bottom Line​The companies that win the next five years of commerce will not be the ones that felt most ready right now. High confidence, right now, is a reason to look harder.​

Forbes Communications Council is an invitation-only community for executives in successful public relations, media strategy, creative and advertising agencies. Do I qualify?