TRUSTED AI ADVISORY INITIATIVE · A SHARED VISION GS1 × PLUGILO
“The Original Returns: to Every AI Answer in the World”
How the real data of the data owners can become the standard for AI advisory. A conversation about a shared vision with Thomas Fell, Lead of GS1 Germany, and Michael Mohr, founder of plugilo and initiator of the Trusted AI Advisory Initiative
GS1 is a neutral, international, not-for-profit organization that develops globally uniform standards for the identification, capture, and exchange of data. More than two million companies are part of GS1, from global brand owners and manufacturers to small businesses with their own products. The GTIN of a product is generally assigned by the brand owner or the owner of the product specification. Through Verified by GS1, basic information on more than 300 million products is available. More than one billion products carry a GS1 barcode; worldwide, GS1 barcodes are scanned around ten billion times per day. Around 120 national GS1 organizations carry these standards into the markets of the world. plugilo is building the trust layer of the AI era: living original data from the owners, directly queryable and distributed to people, partners, and AI systems. At its center is pluggy, the Trusted AI Advisory Agent that will ensure advice based on the real data of the data owners everywhere in the world. In this conversation, both sides explain what the joint proof of concept is meant to demonstrate and why it can become the beginning of a worldwide collaboration that reorganizes the data foundation of AI advisory: away from uncontrolled copies, toward the accountable original of the brands.
Mr. Fell, AI systems answer billions of questions every day, including questions about the products of your customers. What is the problem with that?
Thomas Fell: Today, those answers are largely assembled from a web of copies. Product information is taken over thousands of times by portals, marketplaces, and content systems, shortened, translated, and recombined, usually without source, date, or responsibility. An error is copied three hundred times and then appears to the AI as broadly confirmed consensus, while the manufacturer's corrected original statistically remains a minority opinion. Frequency is mistaken for correctness. Consumers can no longer tell which statement is reliable. Brands, in turn, lose control over the information with which their products are explained, compared, and recommended. And this happens precisely at the point where decisions are made today.
With the GTIN, GS1 has brought order to world trade once before. Do you see parallels?
Thomas Fell: The parallel is striking. More than fifty years ago, every retail chain had its own article identification. What prevailed, with the GTIN, was the one shared number. Today, more than two million companies across more than 25 industries use GS1 standards, and the GTIN is scanned over ten billion times a day. This infrastructure answers the question: which product is this? Through GDSN, companies exchange their master data in a structured way, and with DQX we make data quality measurable. What must be added in the AI era is the final, decisive step: the verified truth of the data owner must arrive where advice is given and decisions are made, meaning in the answers of AI systems, at the shelf, in the shop, in the conversation. That is exactly where I see the starting point for a possible joint undertaking with plugilo.
Mr. Mohr, you call this vision revolutionary. What is genuinely new about it?
Michael Mohr: For the first time, the accountable original itself can become effective in the AI world, instead of being represented there only by uncontrolled copies. That has never existed before. GS1 provides the globally established product identity and the quality-verified owner data base. plugilo turns it into a living digital object: visible to people, machine-readable for systems, directly queryable, distributable everywhere, and permanently connected to the owner's source. GS1 stands for the identity and quality layer of world trade, plugilo for the open trust and distribution layer for AI. When these two infrastructures come together, a path emerges on which the real data of the data owners can reach every answer in the world. The program fits into a single sentence:
| Replace the copy with the original. Everywhere. The chaos of data copies on the internet and in AI systems is superseded by the accountable truth of the owners. |
Before we go deeper: who is plugilo?
Michael Mohr: plugilo is an independent company that emerged from DCI, which has been structuring product and market data since 1993, mapping numbering systems to one another, and connecting data sources. The founding idea, bringing all identifiers of an object onto one shared reference, was our original blueprint, years before the industry formalized it as the knowledge graph. Today, plugilo operates a productive infrastructure of three layers: the owner-controlled master data base, the distribution layer with the plug-it connector, and the application layer with Cards, the app, and the AI advisor pluggy. More than six million prepared data records form the initial foundation. And above it all stands the Trusted AI Advisory Initiative, the open association of data owners that we initiated and orchestrate.
You are working on a joint proof of concept. What is it meant to demonstrate, concretely?
Thomas Fell: Envisaged are real, extensive owner data sets from leading brand manufacturers. Individual records contain well over 100 attributes, images, and documents, GTIN-identified and quality-verified. The proof is meant to demonstrate the entire journey: the GTIN is resolved unambiguously, the released data is transferred completely and correctly into the Master Data Card, provenance and quality evidence remain visible, and pluggy answers real user questions traceably from exactly this base. Just as important: the process must be repeatable and transferable to any number of products and brands. This is not about a lab demo, but about the blueprint for scaling.
Mr. Mohr, what is a Master Data Card, and why do you speak of a “living original”?
Michael Mohr: The Master Data Card is the digital original of a product: all relevant information in one unified object, from attributes, variants, ingredients, and allergens through images, videos, manuals, and certificates to accessories and complementary brand knowledge. For people it is understandable and visual; for machines it is structured and directly processable. Four properties are decisive. It is source-bound: every statement points to its origin and scope. It is versioned: changes remain traceable. It is synchronous: when the owner changes their data, the Card updates automatically everywhere. And it is rights-controlled: use and distribution follow the owner's releases. The released owner version always remains the authoritative source. plugilo holds no competing truth, the Card follows the source. That is why we say “living original”: not a copy, but a permanent connection between brand and human.
How does a consumer experience this, in concrete terms?
Michael Mohr: They stand in front of a frozen pizza, scan the code, and hold the original in their hand. They ask pluggy: Does it contain gluten or lactose? Can it contain traces of nuts? Is it vegetarian? How many calories per serving? How long does it need in the oven? The answers come from the brand's released data, with a visible source. If a piece of information is missing, this is stated clearly instead of being invented. Then comes the step that turns an answer into a relationship: the consumer takes the Card along with one click. It then sits in their plugilo app, the universal watch list: one app for all brands, shops, publishers, and topics, in which people collect, compare, and organize the originals of all providers in one place, in stacks. The cards are alive, so everything always stays up to date. Today, every shop has its own little bookmark heart, and by evening nobody remembers where they saved what. We are standardizing saving the way PayPal once standardized paying.
You call pluggy an “employee”. Why?
Michael Mohr: Because that is the most honest description. pluggy is not a generic chatbot but the best employee a brand ever had: it knows every product, speaks every language, never gets tired, and works in a thousand places at once, always fed from the same released source. It learns from three sides. The brand gives it its knowledge: data sheets, manuals, FAQs, the company's actual advisory expertise. Customers' questions sharpen it every day. And it knows the place where it is currently working: at the retailer its assortment and offers, at the magazine its reviews, at the specialist studio the showroom. The revolutionary part is the consequence: the brand can put this employee at the side of its entire partner network, every retailer, every shop, every portal, free of charge. Precisely where the skilled-labor shortage hurts most, retail gains an ever-improving advisor that explains on-brand, compares performance, and helps sell. Our division of roles is the program itself: pluggy advises. plugilo makes pluggy trustworthy.
| The best employee a brand ever had: in a thousand places at once, in every customer's pocket. And the brand puts it at the side of its entire retail network free of charge. |
And how does this original reach every place on the internet?
Michael Mohr: Through the plug. plug-it is a widget that any shop, portal, or media site can embed without an IT project. Around 80 percent of online shops are reachable via a handful of widespread shop systems. One click on the plug, and the Card, advice included, is with the customer: a new kind of watch list offered as a service to a site's own visitors. Where the plug is not yet embedded, the browser extension takes over. It recognizes brands and products in the text of any suitable web page, and the little plug appears like a new punctuation mark of the internet. One click opens the original data, regardless of who owns the page. Together, an open meta layer emerges above all systems and AI models: source binding, rights and releases, structured queries, and open distribution APIs. On this layer, the real data is omnipresent, available to every user and every serious AI system. The brand no longer has to hope that the AI world represents it correctly. It represents itself correctly. Everywhere.
You speak of a new network under the control of the brands. What does that look like?
Michael Mohr: Let's turn the perspective around: brands are not supplicants of the platforms. They are the owners of the truth and command the strongest reach in the world: their products, their retailers, their customers. Built on established standards and with the plugilo infrastructure, they can, for the first time, convert this brand power into a network of their own. The brand activates its Cards, gives pluggy its knowledge, and spreads both in one stroke across its entire partner network: retailers, shops, and distributors at the point of sale; publishers, portals, and influencers at the point of interest. Each of these places becomes an advisory location with the real data, including for new customers who encounter the brand there for the first time. And every customer who takes a Card along stays connected with brand, retailer, and publisher. In this way, all market participants are connected with end customers in a market-appropriate manner: without noise, without spam, without third-party algorithms, and without the usual distractions of other networks. Owner-controlled instead of algorithm-driven. And the network effect is built in: every brand that joins brings hundreds to thousands of partners with it. Publishers and retail earn along with the advisory instead of being copied and rendered invisible.
Mr. Fell, many claim to be “trustworthy”. How does trust become verifiable here?
Thomas Fell: By not reducing it to a single blanket grade. The model answers three questions separately. First: who is responsible for the data? For this there are five visible trust levels, from structured reference up to permanently managed source truth. Second: how good is the data, measurably? For this there is the Data Quality Profile, covering completeness, currency, consistency, and traceability. Third: what is externally evidenced? For this there is the Evidence Passport with standards, tests, and certificates. Each piece of evidence remains attributed to its issuing organization. Measurable data quality, as delivered by programs like DQX, strengthens the quality profile and the evidence. This keeps it precisely recognizable which statement has been verified and released, and to what extent, for people and AI systems alike. Trust moves from assertion to verifiable property.
What would the individual market participants concretely gain?
Thomas Fell: The remarkable thing about this model is that every participant wins immediately, not only once the network is large. Manufacturers and brands regain control over their binding product truth and scale their advisory across every channel, while service, error, and returns costs go down. Retailers and specialist trade gain an always-available digital advisor and qualified contacts instead of anonymous clicks. Publishers and media become part of the value creation as trustworthy context and earn along. Consumers receive clear answers with a visible source, savable performance comparisons, and one watch list for everything. And the AI systems themselves win: machine-readable, provenance-secured, rights-clear sources. That means fewer false answers and lower liability risks. The formula is simple: maintain once, reliably, and advise trustworthily, worldwide.
What strategic perspective could this open up for GS1, also internationally?
Thomas Fell: The existing identification, exchange, and quality infrastructure would be extended all the way to the place of information, advice, and decision. Manufacturers who have long since done their data and quality work would not have to repeat it in yet another isolated solution, but would activate it for the AI era. The GTIN would then lead directly from identifying to advising. If such a proof shows the substantiated benefit I expect, it can be jointly examined how the model becomes visible and accessible to manufacturers, how associations and technology partners are included, and what role the international GS1 network with its roughly 120 organizations could play in scaling worldwide. The potential is obvious: the building blocks have existed for years; they only had to come together. The concrete design deliberately remains subject to joint evaluation. The target picture, however, would be a long-term, worldwide cooperation.
Mr. Mohr, behind all of this stands the Trusted AI Advisory Initiative. Who should join, and why now?
Michael Mohr: The Initiative is the open, worldwide association of data owners: brands, manufacturers, retailers, publishers, associations, and public institutions. Its charter is immutable and applies to us as the initiator as well: the data belongs to the owner, every release is revocable, every piece of information carries its provenance, the network is open and neutral, and the value created flows to those who create it. Why now? Because the truth layer of AI is being organized in exactly these months, with or without the owners of the data. We invite everyone to organize the influence of real data in the AI era themselves, instead of leaving it to chance and to the AI systems alone. The lever is concentrated: the 1,000 largest brands of the world together reach practically every internet user. Around 50 active supporters are already enough to carry the Initiative into the mainstream. Associations can organize entire industries and protect their members' interests; retail and media actively request the original data of “their” brands. The entry is deliberately low-threshold: the observer level is free of charge and non-binding, one's own presence is secured in minutes, and the benefit begins with the first partner, not only with critical mass.
| The more data owners the Initiative represents, the better it can assert their influence worldwide across all AI systems. The invitation extends to every brand, every retailer, every publisher, and every association. |
Finally: what would success look like a year from now?
Thomas Fell: When an end-to-end proof stands and the first brands have activated their GTIN-identified, quality-verified data as living Master Data Cards: visible, queryable, and distributed via open interfaces. When the process demonstrably scales. And when, on that basis, it can be jointly decided what a rollout looks like: which manufacturers, associations, and international organizations are included in a next phase.
Michael Mohr: And when the market has understood what can emerge here: the truth layer of AI, built by the owners of the data, carried by the reach of the brands, supported by established standards and measurable data quality, open to every serious system in the world. The GTIN brought order to trade more than fifty years ago. Now, the data foundation from which the AI of the world advises can be brought to order. And everyone who joins, wins. The invitation stands: corporate.plugilo.com.
Summary
GS1 Germany and plugilo are working on a joint proof of concept. It examines how GTIN-identified, quality-verified owner data can be activated as living Master Data Cards, used via the AI advisor pluggy for trustworthy, source-bound advice, and distributed through open interfaces to retail, media, and AI systems. The released owner version always remains the authoritative source. Based on the jointly evaluated results, both sides will assess possible steps toward a long-term, worldwide collaboration within the Trusted AI Advisory Initiative.