Skip to content
Engineering made in Germany

If you're investing €30,000, you should know who you're talking to.

DataFlowFacts is an engineering company from Cologne. We don't develop foundation models and we don't run a cloud. We build finished AI systems from open technology that work for mid-sized companies, and put them on your premises. Here you can see who does this, what you get and who your contracting party is.

The people

Who you'll get to know here.

The three people responsible for your project, from the first enquiry to ongoing operation. Behind each of them is a team.

Peyman Javanbakhsh, Solution Manager & Managing Director

Solution Manager & Managing Director

Peyman Javanbakhsh

Responsible for quotes, contracts and the learning programme. Turns customer questions into new episodes for the learning portal and office hours in the webinars.

Get in touch

Florian Schmidt, Customer Success Manager & Head of IT

Team Lead Customer Success & Head of IT

Florian Schmidt

Your dedicated contact person during operation. Leads the support team that knows your configuration and your history. You will not have to explain anything twice.

Team Lead AI Setup & Project Management

Leslie Schäfer

Leads your project with her team from order to handover: clarifies requirements, plans the rollout, connects your knowledge sources and tests the AI against your real questions before you receive it.

The team behind them

Development & support

Behind every team lead is a team: people who set up models, connect knowledge sources, integrate systems, provide support and fill the learning portal. You will get to know them once your project is under way.

We will add Leslie Schäfer's photo as soon as it is available. We do not show stock photos as employees.

The device

What's on your premises.

You're not buying a platform or access, but a system that works on your premises. Three sizes: the hardware is the foundation, the product is the AI running on it.

Packages & pricing
Compact system: quiet desktop or tower computer in the office (product view)

Compact

AI Compact

For the workstation and smaller teams. A quiet desktop or tower computer in the office.

Server class: server in a rack or as a tower in the server room (product view)

Team

AI Team

For multiple users and departments. A server in a rack or as a tower in the server room.

Redundant server class: two systems that back each other up (product view)

Enterprise

AI Enterprise

For larger environments and redundant design. Two systems: if one fails, the other can pick up operation.

Product view of the device. The exact specification is in your quote; we don't name components we don't supply.

Contracting party

Your contracting party is based where your law applies.

Contracts under German law, a company in Cologne, German-speaking contacts who come to you when needed. And no corporate headquarters deciding on terms without ever having seen you.

Contracting party

Werbeeinheit UG (haftungsbeschränkt)

BrandDataFlowFacts
Registered officeChristian-Gau-Straße 17, 50933 Cologne, Germany
ContractsUnder German law, fixed price in writing

Registration details and representation can be found in the legal notice.

What we stand for

Control before convenience.
And three more principles you can hold us to.

Control before convenience.

Every decision that affects your data is yours to make. Updates after your approval, remote access only with your yes, logs in your system. If that sometimes costs one more click, that's the price of sovereignty.

The right size, not the biggest possible.

We won't sell you a data centre when a compact system is enough. The free configurator recommends the cheapest package that covers your needs. That's by design, not an oversight.

Open instead of proprietary.

We rely on open AI models and open standards. You're not tied to us: the system belongs to you, the technology inside it is documented, and you could run it without us. We believe you won't want to, but you could.

Explaining instead of promising.

You won't find a security guarantee on this website, or a “100% GDPR-compliant”. You'll find a description of what the system does and doesn't do. Understanding enables deciding. That matters more to us than a quick deal.

How we work

Technical work, not a sales pitch.

From the first enquiry to ongoing operation, you talk to the people who build and set up the system, not to a hotline.

  • Understand first, then configure. We ask what your staff do, not which model you want.
  • Fixed price, in writing. What's in the quote applies. Additions only happen if you want more.
  • Quality tested on your cases. Before handover, we test the AI against real questions from your day-to-day work, together with you.
  • Progress in real steps. Ten work steps from order to handover, each with feedback to you. No status update without work behind it.

What we don't do

No foundation models of our own

We select, combine and adapt. The foundation models come from open AI research, and we'll tell you which, if you ask.

No cloud platform, no mandatory subscription

We don't run a platform where your data is stored. The system is on your premises. Optional services can be cancelled monthly.

No promises we can't keep

No number of weeks we don't actually know. No guarantee we can't give. We'd rather say an honest “let's discuss that” than give a quick yes.

Ready for your own AI?

Put your package together in a few minutes, or first check whether your existing hardware is sufficient.