Hosting and maintenance

The setup follows how critical the solution is to you, but the foundation is the same.

• Hosting in EU regions or on your own servers, depending on where your data may be processed.
• Monitoring and alerts, so problems are caught before your customers or staff notice them.
• Security updates and dependency updates.
• Model and instruction updates when providers change something, with tests that show the results still hold.
• Cost control per process, so you can see what each agent costs and what it saves.
• A monthly review of what the agents have done and where they can take on more.

I run SökHem and OpenAudit in production myself. SökHem runs on its own servers with Docker, Traefik and Dokploy, with Prometheus and Grafana for metrics and alerting. Those are the same tools and routines I use for client solutions.

So you do not get a consultant who disappears after launch, but someone who has carried the responsibility for systems that need to be up around the clock.

If you would rather run the solution yourselves, your team gets the code, the documentation and the training it needs. Everything lives in your repository from the start and builds on open, common tools.

You can also start with me running it and take over later. There is no lock-in, and the handover is planned from day one.

Maintenance is mostly about what you never see. When everything works it goes unnoticed, and that is the point. But when something happens, such as a provider outage, a model that starts answering differently or an integration that breaks after an update to your business system, I get the alert and fix it.

Every month you get a short summary: how many cases the agents handled, how often a person had to step in, what it cost and which improvement comes next. That makes it easy for leadership to see what the investment returns and to prioritise the next process.