Writing things down keeps you honest. It is harder to be vague on paper than in conversation. So we started this journal: a record of our work, the choices behind it, what we are learning. Not a marketing channel. A place to write what we actually think, when there is something worth writing.
Before the rest of it makes sense, you should know what abi group is and why we built it.
What abi group is
abi group is a small holding company. We are three founders: André Loreth, Anton Brinckmann, and Andreas Neugebauer.
abi consulting is our AI services practice for regulated industries in DACH. We audit AI workflows. We build private AI systems on infrastructure that stays under the client’s control. We run those systems alongside the firms that use them. Family offices, tax consultants, specialty insurance brokers. Industries where data confidentiality is not a slogan, and where trust is earned slowly and lost quickly.
abi services is our product engineering firm. Architecture, development, and operations for teams that are building software. We work mostly with founders and Mittelstand companies who want one partner to take a product from a sketch to a system that runs in production. We like relationships where we are still around when the hard problems surface.
We also build products of our own. evmquery and FDTM are ours, owned and operated inside the group. Alongside them, we run ventures together with external co-founders. INTA Compliance is one of those, an AI compliance system for casinos. The split matters: products are ours end to end; ventures are partnerships where we contribute the engineering and the co-founders bring the domain.
And a small set of open source libraries we maintain, mostly because we depend on them.
Why now
We built a group rather than a single consultancy because the consulting work, the engineering work, and the product work belong in one house. We have watched companies separate these three, and then spend years managing the gaps between them. Ours share standards, share people, and share what each side is learning.
Most companies pick one of the three. We picked all three on purpose, because we believe each makes the others sharper.
What survives a cycle is the engineering. Labels change. The work of running production systems does not.
What we hold to
A few things we have held to since we started.
Engineering over slideware. We have been in rooms where the deliverable was a document and the system it described did not exist. We do not do that. If a deliverable cannot be deployed and observed, we do not bill for it.
Compliance is a feature, not a tax. In regulated work, datenschutzkonform ausgestaltbar is part of the product, designed in from the first sprint. We have seen what happens when it is added at the end.
Boring stack, careful choices. We pick technology we are willing to run in five years. The pull toward something newer is constant. We resist it deliberately.
Small on purpose. Three founders, deliberately. We grow our reach, not our headcount. Every firm that has lost its quality has grown its way out of it.
What this journal is for
This journal is the place we will write things down. Engineering essays, when one of us has built something worth explaining. Product notes, when one of our products changes. Founder reflections, sparingly. Notes on the open source libraries we maintain.
We will publish when we have something worth publishing. Authors are the three of us. The website lists who wrote each piece. We sign our names to what we say.
If something here prompts a question, we are easy to find.
You can follow us on LinkedIn.