Kathrin: Manuel, TAGBASE recently launched its product verification app, but your next move might be even more transformative. Can you tell us about it?
Manuel Mertl: Absolutely. We’re about to launch a feature we call “Chat With Your Product.” It lets users tap a product’s NFC tag, open our platform in their browser — and instead of just verifying the product’s authenticity, they can ask it any question. And the product answers, intelligently and instantly, using AI.
Kathrin: That sounds like something from the future. How are you making that work technically?
Manuel: The key is vector embeddings. Brands upload their product manuals, technical documentation, compatibility charts, and more. We then transform all of that content into vector embeddings stored in a high-performance PostgreSQL vector store. So when a user asks a question, we semantically search the most relevant sections, retrieve them in real time, and use a large language model to craft a precise, context-aware answer.
Kathrin: Why embeddings? What challenges did you face early on?
Manuel: Initially, we just tried passing the full document to the model — but that quickly hit context window limits and performance issues. Our CTO and Co-Founder, Mario Uher, realized we needed a smarter way. He designed a pipeline that chunks, indexes, and embeds documents efficiently. Now we’re loading only what’s relevant to the user’s query — it’s faster, more scalable, and accurate across even massive technical docs.
Kathrin: Can you give us some real-world examples?
Manuel: Definitely:
- Cosmetics – A user taps a serum bottle and asks: “Is this safe to use while pregnant?”
We scan the safety sheet and give an immediate, brand-approved answer. - Spare Parts – A technician scans a brake pad: “Will this fit a 2018 Audi A4?”
The compatibility matrix is embedded, so the answer is instant and precise. - Premium Beverages – Someone asks: “How should I store this tequila after opening?”
The response might include storage tips, ideal serving temperature, or cocktail ideas.
What used to take minutes of searching, downloading PDFs, or calling customer service — now takes seconds with one tap.
Kathrin: Sounds like this goes beyond just verification.
Manuel: Absolutely. TAGBASE started as a product authenticity company — and that’s still our foundation — but in truth, our business is data. We’re a data collection and enrichment machine. Every product interaction, every user question, every tap gives us structured, high-value signals. For brands, that means product insights, usage patterns, geo-location data, and behavioral trends. For us, it means building the world’s most valuable layer of real-world product intelligence.
Kathrin: That’s a bold claim. But from an AI perspective, it’s exciting. How do you see this evolving?
Manuel: We see the future of physical products as intelligent, interactive, and trackable. Each verified product becomes its own data node in a global network — secured by NFC, enriched with AI and optionally verified by blockchain.
Long-term, we’ll personalize the experience: imagine the same question answered differently based on user location, previous interactions, or even device type. We’re building infrastructure for that world now.
Kathrin: And for investors?
Manuel: Investors looking at AI should understand: this isn’t just another chatbot. This is AI meeting physical infrastructure. We’ve solved authentication, we’re solving support, and we’re unlocking a data layer most brands don’t even know they’re missing.
#AI #VectorSearch #NFC #CustomerExperience #DataCompany #TAGBASE #ChatWithYourProduct #InvestInAI #Startup #SmartProducts