Even before a packaging buyer sends out an inquiry, artificial intelligence may have already compared suppliers, evaluated material structures, and formed price expectations. A study conducted by info.link specifically for INNO-Pouch 2026 shows which companies become visible in this process – and why technological expertise alone is no longer enough.
Anyone searching for a recyclable stand-up pouch today may no longer begin their research on Google or at a trade fair, but with a prompt to ChatGPT, Perplexity, or another AI assistant.
Which manufacturers offer recyclable mono-material pouches? What barrier properties can be achieved? Can the structures be processed on existing packaging machines? And how much more expensive will the transition from PET/aluminum/PE to mono-PE or mono-PP be?
These are classic questions that were previously clarified in direct discussions between packaging developers, technical sales representatives, and buyers. Increasingly, however, they are being answered even before the potential customer makes contact with a supplier.
The problem: AI does not necessarily rely on the companies with the greatest technical expertise. Rather, it relies on those whose knowledge it can find and process on the internet.
Chris Goeken, co-founder of info.link, demonstrated at INNO-Pouch 2026 in Barcelona just how wide this gap already is. Specifically for the event, his company examined which sources AI systems consult when answering typical questions about purchasing flexible packaging.
The results are likely to be surprising for many companies in the FlexPack industry.
665 AI Answers – and a Remarkable Gap
For the study “How AI is changing the Pouch Purchasing Process”, info.link examined five typical questions from the stand-up pouch procurement process over a 35-day period, from September 7 to October 11, 2026.
The English-language prompts covered supplier search, recyclable high-barrier structures, spout packaging, processing mono-materials on existing machinery, and the costs of a material transition.
Across four AI systems, including ChatGPT, Perplexity, and Google AI Mode, a total of 665 answers with 5,748 source citations were evaluated.
<text strong=”true”>The main findings of the info.link study:</text>
- Only 13 of the around 80 companies represented at INNO-Pouch were cited as a source at all.
- Only 6.6 % of all source references accounted for these companies.
- Mondi was cited 199 times – more often than all other represented companies combined.
- Seven out of nine suppliers in a sample ChatGPT shortlist were found via a specifically named product line.
- A single information page from a label specialist achieved comparable visibility to the entire Gualapack website.
These figures do not mean that the remaining companies offer no relevant solutions. Rather, they demonstrate that a significant gap can exist between existing technical expertise and digital visibility.
Goeken framed the challenge in his presentation along these lines: Companies invest a lot of money in material development, machinery, and new packaging solutions. But are they also investing in ensuring that AI systems understand these innovations in the first place?
Why a Single Good Webpage Can Overtake Major Brands
One of the most striking findings concerns the importance of individual technical information pages.
The most frequently cited single page in the entire study did not come from a major film manufacturer or packaging converter, but from LabelProfi, a label specialist.
Its information page on PPWR compliance for stand-up pouches was cited 106 times as a source.
The entire Gualapack website achieved a similar presence in the analyzed responses. Amcor, by contrast, was cited as a source in only about 5 % of responses, and Huhtamaki in about 2 %.
Why?
The explanation apparently lies less in the size of the company than in the way information is provided.
A clearly structured page that fully answers a specific question is easier for AI systems to use as a source than an extensive corporate website where the same information is scattered across press releases, product brochures, and PDF documents.
Mondi demonstrates how technical expertise can be made discoverable in a targeted manner. The company has dedicated pages for mono-material stand-up pouches, high-barrier structures, and retortable mono-material packaging. The content focuses on specific applications and questions.
In addition, there is independent validation in the form of certifications, customer projects, and technical publications.
For small and medium-sized packaging companies, this represents a real opportunity. They do not necessarily need the largest marketing budget to be included in AI responses.
The decisive factor will be whether they answer their potential customers’ questions better than others.
AI is Already Creating Supplier Lists – Including Technical Specifications
The development becomes particularly interesting where AI systems do not just summarize information, but prepare concrete procurement processes.
When asked for leading European manufacturers of recyclable mono-material stand-up pouches, ChatGPT put together a list of nine suppliers on October 4, 2026.
These included Mondi, Amcor, Coveris, Gualapack, Südpack, Wipak, Huhtamaki, ProAmpac, and Constantia Flexibles.
But the response did not end there.
ChatGPT additionally offered to prepare a concrete request for quotation (RFQ). This included a technical checklist with requirements for layer structure, material proportions, oxygen and water vapor barrier, proof of recyclability, closure materials, and machine trials.
This is a remarkable development for flexible packaging sales.
Previously, technical sales representatives could explain in personal conversations why a specific material structure makes sense, which properties are actually needed, and which requirements might be excessive.
If the AI formulates a specification even before this conversation takes place, it shapes the buyer’s expectations.
This can be helpful because customers enter technical discussions better prepared. However, it can also become problematic if unsuitable requirements or inaccurate assumptions are adopted.
It was also striking that seven of the nine suggested companies were named via specific product lines, such as AmPrima, re/cycle, or PurePE/PurePP.
A clearly named solution with its own product page seems to make it easier for AI systems to assign technical properties to a specific offering.
For packaging manufacturers, this raises a new question: Is it still sufficient to provide general information about sustainable packaging on the website? Or must individual solutions be described with their technical properties much more precisely?
When Manufacturers Do NOT Quote Prices, AI Sets the Price Anchor
Another finding of the study is likely to particularly concern sales managers.
info.link examined how AI systems assess the costs of converting from a conventional PET/aluminum/PE laminate to recyclable mono-material structures.
On September 29, Perplexity cited a cost markup of 15 to 25 %.
Just six days later, the response from the same system was 3 to 8 %.
The underlying sources were predominantly from information pages and general cost overviews. None of the responses analysed cited European packaging manufacturers as sources.
In one case, the AI even quoted an obviously implausible packaging price of 1.33 to 1.61 US dollars per 1,000 pouches.
The example demonstrates a fundamental problem: AI systems often deliver specific numbers even when the available database is insufficient.
This can create difficult situations for sales.
A buyer who has previously read about a cost markup of only 3 to 8 % may critically question a significantly higher calculation. Another customer expecting 25 % extra costs might walk away from a project even before submitting an initial inquiry.
Yet actual costs depend on numerous factors: barrier requirements, material structure, production volume, printing, spouts, zippers, and processability on existing equipment.
The study therefore does not recommend publishing confidential price lists. Rather, manufacturers could explain typical cost ranges and the key influencing factors.
Those who do not explain the economic contexts of their products themselves may leave this classification to sources that know the market significantly less well.
From Search Engine Marketing to Visibility in AI Responses
The consequences go beyond traditional search engine optimization.
Goeken reported on a project with the hair care brand Guhl, in which the website was specifically tailored to provide product information in a format easily understandable for AI systems as well.
To this end, numerous specific questions about the products were answered. The answers are structured, referenced with sources, and regularly checked and updated.
According to Goeken, more than 500 such question-and-answer items were created for Guhl, updated monthly. In the competitive environment studied, the brand achieved a strong presence in AI responses.
info.link worked with a similar approach at a German packaging company in the corrugated board sector as well.
Applying this approach to flexible packaging is a logical step.
A film manufacturer could, for example, answer specific questions: What OTR values can be achieved with a specific mono-PE structure? Under what conditions is packaging recyclable? What sealing temperatures are recommended? How does a spout alter the recyclability assessment?
Such information has long been available in many companies. However, it is often locked away in technical data sheets, internal presentations, or documents that are only made available upon personal request.
Goeken’s approach is to organise this existing knowledge in such a way that it can also be found, understood, and cited by AI systems.
This is not about publishing as many generic texts as possible. The key is technically sound answers to specific questions that potential customers actually ask.
What Manufacturers of Flexible Packaging Can Do Now
info.link derives eight recommendations for action from the study. In practice, these can be condensed into four key tasks.
First: Check your own visibility. Companies should work together with their sales team to gather typical questions from potential customers and investigate them across various AI systems over several weeks. Who is mentioned? Which sources are cited? Which competitors appear regularly?
Second: Make technical information more accessible. Material structures, barrier values, sealing properties, and proof of recyclability should not exist solely in PDF data sheets or behind contact forms. Well-structured websites with concrete technical details can improve discoverability.
Third: Describe solutions instead of general corporate promises. The results show that clearly named product lines with dedicated information pages frequently appear on AI-generated supplier lists.
Fourth: Back up claims with independent sources. Certifications, verifiable customer projects, and publications in trade media help make technical statements verifiable.
This is not a guarantee for more frequent AI mentions. Responses change constantly, and different questions can produce completely different results.
Nevertheless, the study shows that companies do not have to leave their digital visibility entirely to chance.
And What Does This Mean for Innoform?
Innoform was also included in the study.
With ten source references, Innoform was among the 13 companies represented that appeared at all in the AI responses examined.
While gratifying on the one hand, this also demonstrates the immense potential for growth even for companies that already publish extensive technical information.
For Innoform, this development is particularly interesting because our content is not primarily intended to sell individual packaging products. We want to convey technical knowledge, contextualize developments, and connect experts along the value chain.
It is precisely this type of information that can play a key role in AI-supported research.
The challenge lies in the fact that technical expertise alone does not automatically lead to digital visibility. Knowledge must be structured and accessible in a way that allows modern information systems to process it reliably.
Conclusion: The Competition for Customers Begins Earlier
The study by info.link is not a representative survey of the entire European packaging market. It is based on five English-language questions analyzed over 35 days with country settings set to Germany. Different questions, languages, or applications could lead to different results.
Nor is a source citation by an AI equivalent to a supplier recommendation.
Nevertheless, the investigation provides an important indication of how industrial procurement processes could evolve.
The competition for customers may well begin long before sales learns of a specific project.
This creates an additional task for flexible packaging manufacturers. Alongside material development, processing technology, and regulatory compliance, the digital discoverability of technical expertise is becoming increasingly relevant.
Those who develop innovative solutions should therefore also ensure that potential customers—and the AI systems they use—can find and correctly interpret these solutions.
The crucial question is not whether AI will control every packaging purchase in the future. Rather, it is whether companies are prepared when their customers enter the initial meeting today carrying an AI-generated supplier list.
Have you ever checked which manufacturers ChatGPT or Perplexity recommend when asked about your packaging applications? And does your own company appear in those recommendations?
About the study: “How AI is changing the Pouch Purchasing Process” was created by <Link url=”https://info.link” title=”info.link”/> under the direction of Chris Goeken for the participants of INNO-Pouch 2026 in Barcelona. The study includes 665 AI responses and 5,748 source citations from the period from September 7 to October 11, 2026. This article is based on the complete study as well as the transcript of Chris Goeken’s presentation at INNO-Pouch.
