The approach
Start with the business decision, not the technology.
The useful question may be where work repeats, how leaders should prioritise ideas or which architecture is justified. The technology follows that context.
About Sobrn
An independent AI and automation practice for European businesses.
Sobrn helps leadership teams choose practical AI priorities, turns repeated processes into reliable workflows and reviews the architecture behind consequential AI systems.
The outcome
Advice and systems your team can use.
Recommendations, scope, ownership, exceptions and access are made explicit. Written outputs and handover are part of delivery rather than an afterthought.
Working principles
Five things every project should make clear.
-
1
Start with the process. A rule, API call, better form or removed step may be more useful than AI.
-
2
Agree the scope first. Inputs, ownership, exceptions and outcomes are clear before implementation begins.
-
3
Keep people in control. Consequential decisions and uncertain cases have explicit human checkpoints.
-
4
Stay vendor-neutral. Tools and model providers are selected for the workflow, without platform incentives.
-
5
Keep your systems yours. Customer-owned accounts and infrastructure are used wherever practical.
Founder
Rui Abreu
10+ years across software engineering, machine learning and data systems, including production NLP and cloud infrastructure. Sobrn applies that experience across leadership training, workflow implementation and architecture decisions for European businesses.
Scope and geography
EU-based and intentionally focused.
Sobrn works with EU and EEA businesses on clearly bounded leadership workshops, workflow automation projects and architecture reviews. Data sensitivity, supplier location and human control are considered in proportion to the engagement.
For a first conversation, describe the decision or workflow through the project form or email ruiabreu@sobrn.eu.
Start with the question in front of you.
A few sentences about the team, workflow or architecture decision are enough. You do not need to know the final scope yet.