If you have read the openProd Academy, you have met Ingrid Strand, Zofia Kamińska and Nate Holloway. They sign most of our articles and they present our films. None of them is a real person.
This post explains who they are, why we work this way, and what stays human.
Who they are
Three AI characters, each with a topic area:
- Ingrid Strand covers enterprise data strategy: what product data work costs, where the market is going and how PIM vendors compare. In our films she plays a Head of Product Data.
- Zofia Kamińska covers the hands-on side: supplier files, onboarding workflows and day-to-day catalogue work. In the films she plays a Product Data Manager.
- Nate Holloway covers the technical side: integration architecture, APIs, MCP and the protocols agents use. In the films he plays a CTO.
You can see all three in our launch film, Introducing openProd.io: every supplier file, one catalogue. Their faces and voices are generated with AI tools.
They are labelled wherever they appear
On every article they sign, an “AI character” label sits next to the name, and the byline names the person who reviewed the article. Their author pages say in the first line that they are AI characters created by openProd, not real people. The RSS feed adds “(AI character)” to their names. In the structured data that search engines read, their articles are attributed to openProd as an organisation, with me as the reviewer, not to a fictional person.
The characters have no experience of their own, so they should not claim any. The rule for their articles is simple: no “in my ten years of projects” under Zofia’s name. Where an article draws on implementation experience, it says whose it is: LemonMind, the Pimcore partner behind openProd, which has implemented PIM and e-commerce systems since 2008. Posts about my own work, about events we attend and about company news carry my name, like this one.
Every article they sign is reviewed by me
The text under their names is written with AI. Before it is published, I review it. My job in that review is to check the product facts and the numbers, and to make sure the article says something we can stand behind. If you find an error in an article signed by one of them, the mistake is mine as the reviewer, not theirs.
Why a small team works this way
When we introduced the characters on LinkedIn, I wrote: “Meet our freshly hired team of AI experts… Yes, they come ‘from the future’ of AI and are not real people. But the challenges they describe are very real. And our product tutorials get even more real: they are recorded in the actual system, on real import runs.”
That is still the reason. We are a small team. The knowledge behind the Academy comes from years of PIM and supplier data projects, but writing it up regularly, for very different readers, is a lot of work for a few people who also build the product. The characters give each topic a consistent voice: Ingrid for the person who owns the budget, Zofia for the person who opens the supplier file, Nate for the person who connects the systems. AI does much of the writing. People make the decisions.
What I did not want was a byline that pretends: a stock photo and an invented human name on text written with AI. If we use characters, readers should know they are characters.
What stays human
- Decisions. What we write about, what openProd does and what we recommend.
- Product facts. Features, workflows and numbers come from the product and from measured runs, not from a model’s imagination. The onboarding cost calculator, for example, states which of its figures are measured and which are planning assumptions.
- Review. Every article a character signs is reviewed by me before it goes live.
- Tutorials. Our product tutorials are recorded in the actual openProd app, on real import runs. An AI character may present them. What you see on screen is the real system.
Why we say it out loud
I think AI-generated or AI-assisted text published for the public should be disclosed. Readers judge a text partly by who wrote it. If a model wrote it, they should know, and they should know which person stands behind it. That is why every byline on an article written by a character carries the “AI character” label and a named human reviewer.
The EU AI Act’s transparency rules point the same way. Article 50 of the AI Act deals with transparency for AI-generated content, and for text published to inform the public it ties the question of disclosure to human review and to someone holding editorial responsibility. I will not tell you what the regulation requires of your company; that is a question for your lawyer. We label the characters because it is the honest thing to do.
Meet them
- Our AI guides, with the three characters and the reviewer
- All Academy authors
- Product tutorials, recorded in the real app
