Interview with Stavros Kyriakidis, Co-Founder & CEO - Pyx AI

My background is in enterprise software, digitalization and AI. With Pyx AI, we chose to focus very deeply on material compliance in packaging - an area where highly qualified experts still spend a great deal of time reading supplier declarations, specifications and test reports manually. Today, my role is to connect that packaging expertise with technology and turn complex compliance workflows into scalable software.

One of the biggest challenges is that packaging companies are being asked to make faster, better-documented decisions while the underlying information remains fragmented across suppliers, documents and internal systems. The industry needs more standardized, machine-readable material data and much tighter links between regulatory affairs, sustainability, procurement and product development. PDFs will remain important evidence, but they should no longer be the workflow.

We want Pyx AI to become part of the digital infrastructure behind material compliance. A company should be able to understand the regulatory status of its materials, the evidence behind it, open information gaps and where those materials are used. When a supplier document or requirement changes, the affected materials and packaging structures should become visible immediately.

What excites me most is the move from reactive compliance to continuous compliance. Instead of starting a new research exercise every time a customer, supplier or regulation changes something, companies should already know what is affected and where action is required. That shift from searching for answers to managing a living body of material knowledge is a major opportunity.

First, start with the actual application and its constraints. Second, treat performance, sustainability and compliance as one system rather than separate topics. Third, make important decisions traceable - especially when materials, suppliers or requirements change later.

Technical performance remains fundamental, but responsiveness and transparency are becoming just as important. Customers increasingly expect fast, reliable answers about food contact, substances, recyclability, recycled content and regulatory requirements. Suppliers that can provide those answers quickly, consistently and with evidence become easier to work with - and that is a real competitive advantage.

The hardest part is having to optimize several things at once: performance, cost, recyclability, material use, compliance and speed to market. Those factors are tightly connected. Change an adhesive, coating or polymer to improve one dimension and you may immediately create new questions in another. The challenge is not just finding an alternative - it is understanding the full consequence of the change.

The most transformative shift is that sustainability is becoming a measurable design constraint rather than a broad ambition. At the same time, digitalization and AI are making it realistic to manage much more material, regulatory and production data. That combination is changing how packaging is designed, documented and improved.

I expect packaging to become more circular, more traceable and more data-driven. Material efficiency, recyclability and recycled content will increasingly be designed in from the start. At the same time, environmental and regulatory claims will face greater scrutiny, so companies will need stronger evidence behind them. The physical package and the information behind it will become much more closely connected.

I would enjoy seeing more 'what happened next?' stories. There are many announcements around new materials, recycling technologies, automation and AI, but the really valuable lessons often come later: What worked in production? What failed to scale? What measurable result was achieved? Those follow-up stories can be incredibly useful for the industry.

AI will shorten the distance between a complex question and a good decision. In design, it can help compare concepts, materials and historical performance. In production, it can support quality control, predictive maintenance and process optimization. In supply chains, it can improve forecasting, supplier evaluation and risk detection. The biggest impact will come when AI is embedded into real workflows rather than used as a standalone tool.

Data-driven packaging can improve traceability, condition monitoring, logistics and consumer information. But the value is not only in sensors or QR codes. Better data about materials, performance and product use can also help companies make faster decisions in development, procurement, quality and sustainability. The key is turning data into decisions rather than simply collecting more of it.

Automation and robotics will continue to improve speed, consistency and flexibility, especially as manufacturers handle more variants and shorter runs. What is interesting is the imbalance between highly automated production lines and many still-manual surrounding processes. That gap will close: the next automation wave will increasingly include quality, documentation, planning and other information-heavy work around production.

Our role at Pyx AI is on the material-information and compliance side. We do not develop the packaging material itself; we help companies evaluate changes such as recycled content, mono-material structures, new coatings or alternative adhesives against the relevant regulatory and supplier evidence. That helps sustainability decisions move faster without treating compliance as an afterthought.

Packaging will increasingly connect the physical product with digital information. That creates opportunities for personalization, product guidance, provenance, recycling instructions and richer brand stories. But credibility will matter more than simply adding information. The strongest engagement will come from claims and stories that are genuinely useful and can be substantiated.

The next wave of innovation will require closer collaboration between material science, packaging engineering, brands, recyclers, regulatory experts and digital technology providers. Packaging is a system: a material can be technically excellent and still fail if it cannot be processed efficiently, documented properly, used in the intended application or handled in existing recycling infrastructure. The best solutions will optimize across those boundaries.

Spend time understanding the industry before trying to change it. Packaging looks simple from the outside, but the deeper you go, the more interconnected the technical, regulatory and commercial constraints become. For startups especially, start with a specific painful workflow rather than with the technology. AI will become increasingly accessible; deep domain understanding will not.

I do not have one single mentor. I am most inspired by people who combine deep expertise with curiosity and the willingness to question established ways of working. That combination - knowing a field very well without assuming that its current processes are automatically the best ones - is something I value a lot.

I do not have an elaborate morning routine. I try to identify the one or two things that would genuinely move the company forward before the day becomes reactive. In a startup, being busy is easy. Choosing the right problem is harder.

I try to minimize context switching, work with concrete examples whenever possible and make decisions explicit. One real customer problem or edge case usually teaches you more than a long theoretical discussion, and written decisions prevent the same questions from being reopened again and again.

I enjoy turning complexity into something clear and useful. I like problems where the underlying subject is genuinely difficult, but the person using the result should not have to experience all of that complexity. A principle I come back to is: complexity may be unavoidable in the domain, but it does not have to be unavoidable in the process.