Automation examples
- Syncing data between CRM, accounting, shop and spreadsheets automatically instead of retyping it.
- Classifying incoming emails, documents or requests, extracting information and routing it to the right place.
- Generating and distributing reports automatically.
- Turning notifications and approvals that currently happen by shouting across the room or via spreadsheets into a traceable workflow.
AI agents with a sense of proportion
Large language models can understand unstructured text, summarise it and prepare decisions. Making that reliable in a company takes more than a prompt: clear limits on what an agent may do, human approval at critical points, logging, testing with real data and a plan for failure cases.
Philipp Kohnen uses AI where it measurably saves work – and classic automation where a simple rule is more reliable and cheaper.
Data protection and data sovereignty
With sensitive data it matters where a model runs and who sees the data. Depending on the requirements, European AI providers, self-hosted open-source models or a combination are used.
Frequently asked questions
Which processes are suitable for automation?
Mainly workflows that happen often, regularly and in a similar pattern – such as moving data between systems, processing documents or recurring reports.
Can AI agents be used in a GDPR-compliant way?
Yes, if the provider, hosting location and data flows are chosen deliberately. Options include European providers or self-hosted open-source models.
What happens when the AI makes a mistake?
Critical steps get human approval, decisions are logged and failure cases are planned for from the start.
More services
Contact
Want to talk about a project? Philipp Kohnen is happy to hear from you.
Philipp Kohnen · Berlin, Germany
Email: mail@kohnen.digital
Phone: +49 160 90 22 88 20