We build the most complete product data layer for agentic commerce.

It started as an agency problem.

For more than a decade, we built, scaled and ran e-commerce: our own brands alongside client work at sixa AG, the Swiss digital enablement agency for brands, webshops and marketplace sellers. Brand stores, marketplace operations, B2B shops and multichannel rollouts, in electronics, kitchen, fashion, sports and industrial parts.

Every project ran into the same fight over product data. Supplier files arrived half-empty. Every channel wanted a different shape. Channels that could have sold more stayed switched off, because nobody had weeks to map attributes. And every time, the team that gave the buyer something factual to work with won.

The team that invested in product content always won. The hard part was that it was always the hardest thing to do well.Ian Schäfer

Then the buyer changed.

AI assistants now do part of the shopping. They know their users: the bike, the intercom, the budget. They turn a vague question into exact needs and recommend only products whose data confirms every one. Shopping agents filter on attributes, AI search quotes product pages, and Google and marketplaces show complete listings for more searches.

A thin product record is invisible to all of them. Generating more copy does not help either: AI writes only as well as the facts underneath, and those facts were rarely there.

So we built the data layer we always wanted.

Central keeps one verified record for every product. Its agents read what you already have, your shop, your PIM, spreadsheets and supplier feeds, as they are, without a mapping project. Then they research the rest the way your best category expert would: on manufacturer pages, at specialist retailers, in independent reviews and in verified databases.

They keep only what independent sources prove. Every value carries its sources and a confidence score, conflicts go to your team, and your own data is never overwritten. The result is the most complete product data your catalog has had, ready for people, search engines and AI assistants alike.

Connected to every channel.

From that one record, Central writes the content every product needs, in every language your markets speak, and shapes it for each channel. Titles, attributes, limits and rules are built in, and every listing is checked before it ships. Your shop and PIM get the finished content back, and when a fact changes, every text that uses it is flagged for a refresh.

Turned into grounded, on-brand marketing.

Complete data is the first half of our vision. The second half is everything you create from it. Studio turns the verified record into studio shots, gallery images and marketing copy at catalog scale, in your brand's voice and shaped for each channel, and every callout is a fact from the record. Video is coming soon.

What we believe.

• AI drafts. Verification decides. Models are useful writers. What gets published is decided by sources, agreement and your team.
• Only verified claims ship. A claim no source supports is held back with its evidence, so nobody publishes a guess.
• Agents work. People decide. Central runs the work end to end. Your team sees only what needs a human.
• Your data stays yours. Central adds, proves and shows its sources. It never overwrites what you own.

Our manifesto explains why we hold to these rules.

Who we are.

Central is an independent company, built in Zurich. We started it in late 2025 to build what we had wished for in every project, and sixa AG is where we ran the experiments that showed it had to exist. If your catalog has a problem you want fixed first, tell us about it.