Most industrial manufacturers have made real investments in product technology. But having the right tools is not the same as having a working system.
That is the central finding of “The PXM Gap,” a study of 100 senior industrial manufacturing leaders we commissioned with Inriver, Bynder, and Worldwide Business Research. The study asked a straightforward question: Do most manufacturers have the product content infrastructure to keep up with customer expectations and the rapid advance of AI? The answer, in most cases, is not yet.
The tools exist. The integration does not.
Of respondents, 68% have a PIM system in place, and 58% have a DAM system. Half use a product comparison or recommendation engine. By conventional measures, these organizations are invested in PXM (product experience management).
But when asked how product information flows between systems, only 7% describe it as fully integrated. Another 63% say most systems are connected, but manual work is still required to fill the gaps. And 27% say only a few key systems are connected, leaving most processes manual.
The resulting tech stack may look complete from the outside but requires constant manual intervention from the inside. Teams that are effective at keeping product content accurate (43%) are still not effective at moving it quickly. Only 13% say they are very effective at reducing manual effort in moving product data between internal systems.
“Only 7% of manufacturers describe product information flow as fully integrated across all systems. The other 93% are managing some degree of manual process to fill the gaps.”
The PXM Gap, 2026
WORLDWIDE BUSINESS RESEARCH, NTARA, INRIVER, BYNDER
Where it becomes urgent: AI
This integration gap creates channel inconsistency and requires skilled team members to do work that should be automated. And when AI enters the picture, it becomes a strategic liability.
Nearly 80% of the manufacturers in this study have an AI strategy and have already started implementing it. Their top AI use cases all depend on product data that is accurate, well-structured, and accessible across systems:
- Structuring product data so AI tools can read it more effectively (54%)
- AI-assisted product discovery for customers (43%)
- Automated syndication of product content to channels (40%)
None of those use cases can work on a fragmented infrastructure. A manufacturer working to structure product content for AI tools cannot do so effectively if that content lives in disconnected systems. AI-assisted discovery returns poor results when it draws content inconsistently across channels. Automated syndication cannot function reliably without the integration layer to support it.
The 12-to-1 gap
The study’s most striking finding came from segmenting respondents by AI maturity and comparing integration maturity within each segment.
Only 11% of respondents call their AI implementation mature, and 36% of them also have fully integrated product data. Among the 68% still early in AI adoption, that drops to just 3%. That’s a 12-to-1 gap.
But this does not prove causation. Some businesses with mature AI practices reported partial integration. This suggests that they’ve established workarounds or their use cases don’t require full integration. But the pattern across the dataset is consistent: tighter integration correlates with further progress in AI.
The barriers are structural
Closing the integration gap is not primarily a priority problem. The barriers most manufacturing leaders reported in this study are structural.
- 52% cite legacy technology that is difficult to connect with modern systems as a top barrier.
- 47% point to too many tools with overlapping or conflicting functionality.
- 46% say they lack the internal skills or expertise to manage integrations.
These three barriers reinforce each other. Organizations that accumulated point solutions over time often end up with a stack that is difficult to rationalize and even harder to connect. Without in-house integration expertise, connecting or replacing systems can be an uphill battle. And budget pressure compounds it all.
What the AI-ready 11% did first
Manufacturers that have made progress on AI maturity did not build integration as a separate initiative. They treated PXM (the full integration of PIM and DAM at the center of the product content ecosystem) as the foundation from which AI use cases could be layered.
PXM is not a platform. It is the operating model that governs how product content moves between systems. It dictates how they are enriched and validated, and how they reliably reach channels. When a proper PXM foundation is in place, AI tools have consistent, accurate inputs. When it isn’t, AI tools amplify the same data problems that already exist.
Manufacturers that build a true PXM foundation now will arrive at AI-driven commerce with data that works. Those that continue to work around disconnected systems will find that their AI tools surface the same data problems they already have, but at a much greater scale and speed.
The full research report, including the complete segmented analysis and suggestions for building the foundation AI-driven commerce requires, is available free here.