AI and Printing: Why Industry Professionals Want to Influence the UCP Protocol
AI assistants are beginning to change online shopping experiences. But ordering printed materials requires a grasp of factors that aren?t present in traditional retail. Three European trade associations want to incorporate these considerations into a common protocol.
Initiative Online Print (IOP), the German federation Bundesverband Druck und Medien (BVDM), and Intergraf, which represents the interests of the European graphic arts industry, are working on an extension of the Universal Commerce Protocol (UCP). Their goal is to enable artificial intelligence agents to configure and order printed materials while adhering to manufacturing constraints, from file verification through delivery. This approach also raises a commercial question: Who will set the rules for transactions when the customer uses an AI assistant rather than a printer?s website?
With UCP, Google is paving the way for AI-driven commerce
The UCP protocol is designed to enable AI assistants to communicate directly with vendors' sales systems.
On January 11, 2026, at the NRF trade show in New York, Google unveiled the Universal Commerce Protocol. Developed in collaboration with several e-commerce players, including Shopify, Etsy, Wayfair, Target, and Walmart, this open standard aims to facilitate transactions between artificial intelligence agents, merchants, and payment providers.
The concept involves allowing a consumer to express a need to an AI assistant and then complete the purchase within that conversational environment. The protocol defines the information and operations that systems must exchange to identify an offer, process an order, complete the payment, and track the order.
Why Printed Products Defy Traditional Business Models
UCP was originally designed for catalog items. A product typically has a part number, identified variants, a price, and shipping terms. Print-on-demand operates on a different principle.
A print order cannot be reduced to just a product code, quantity, and price. A 16-page brochure can be produced in several formats, using different paper types, weights, binding methods, and finishes. The print run, printing process, page layout, and finishing requirements all affect the production cost. The combination of these factors often requires a custom price quote.
File handling is a second difference. The buyer must provide files that are ready for production, typically PDFs prepared according to the printer?s specifications. These files undergo a prepress check that focuses on bleed, dimensions, fonts, image resolution, and color specifications.
Added to this is the approval of the final proof (BAT). In many workflows, this step determines whether production can begin. It requires explicit approval, and the date and content of that approval must be traceable.
Finally, the delivery time depends on the workshop?s workload, the availability of materials, the options for amalgamation, finishing processes, and shipping. It therefore does not always correspond to a fixed lead time associated with a catalog item number.
A sector-specific extension to incorporate manufacturing constraints
The roadmap identifies seven key functions. These include product configuration and dynamic pricing, file transmission and preflighting, proofing and approval management, production lead time calculation, order rules specific to customized products, production tracking, and restocking management based on saved configurations.
The extension also supports additional modules. These cover environmental data, FSC and PEFC certifications, regulatory compliance, variable data printing, and B2B pricing terms.
Integration with MIS and prepress systems will be critical
In order for an AI agent to place an order for a printed item, the sales information must be linked to the shop floor's technical data.
Creating a common language is not enough. The printers' computer systems must also be capable of providing the requested information and performing the corresponding operations.
MIS and ERP software play a role here at several levels. In particular, they centralize pricing rules, bills of materials, orders, customer data, and certain planning information.
Prepress systems, for their part, handle the receipt, verification, and preparation of files. The JDF and XJDF workflows, developed through the CIP4 initiative, enable the exchange of structured information about jobs and production operations.
UCP does not replace these tools. It serves as a business exchange layer that must be able to communicate with them.
In a possible architecture, the AI agent sends a structured request to a UCP service. The UCP service queries the configurator and the quoting engine, then retrieves the necessary information from the MIS. After confirmation, the files are routed to prepress, and the order data is fed into the production management systems.
This structure requires a precise alignment between commercial attributes and industrial parameters.
Paper described as ?recycled? in a conversation does not, on its own, constitute a reference that a production facility can use. The system must know the available basis weights, sheet sizes, any required certifications, machine-processing constraints, and substitution rules.
The reliability of delivery times is another critical issue. To automatically respond to an urgent request, the sales department must have sufficiently up-to-date information on production capacity. A delivery date calculated without taking into account processing time or transportation availability may result in a commitment that cannot be met.
Business automation will therefore not eliminate the work involved in data preparation. It will simply make it more visible.
Standardization also raises the issue of commercial control
When the AI agent acts as an intermediary in the purchase, data quality and the terms of access to offers take on new importance.
The business challenge lies elsewhere. If an assistant can compare multiple printers and process orders directly, the criteria used to select suppliers become critical.
The cost, lead time, and availability of manufacturing options can be easily processed by a machine when they are structured. On the other hand, technical expertise, the quality of customer support, and the ability to manage a complex project are more difficult to translate into standardized metrics.
The risk is that competition will focus on criteria that are immediately comparable. Printers will therefore have to decide what data to disclose, under what conditions, and with what level of precision.
This issue is particularly relevant to B2B relationships. Contractual rates, volume discounts, payment terms, and catalogs reserved for certain customers cannot be treated as public prices. The future protocol must allow for the differentiation of offers based on access rights and business relationships.
Legal liability is another issue. Article 16(c) of European Directive 2011/83/EU provides for an exception to the right of withdrawal for goods made to the consumer?s specifications or clearly personalized. This exception does not exempt the seller from its other legal obligations. In a process managed by an AI agent, it will be particularly important to ensure that the consumer receives accurate information and to retain the necessary evidence for the transaction.
The project's authors also acknowledge the risk of dependence on platforms. If the customer places an order through a third-party platform, the printer can retain control over production and commercial responsibility while losing some control over the purchasing interface.
Bernd Zipper, president of Initiative Online Print, summarizes the governance issue as follows:
"Those who set the standard set the rules of the game."
This statement raises a question: What role do French printers and their representative organizations play in the development of this future standard? While two German organizations are behind the project, alongside Intergraf, the absence of French stakeholders among the initiators is striking.
Especially since the technical choices made could, in the long run, determine how printers gain access to orders generated by AI agents. It remains to be seen whether French trade associations intend to participate in the standardization efforts and influence the protocol?s direction.
