I’m one of the developers behind Luther. We make it easy to operate “mega-workflows” – the long, cross-team, multi-system processes (insurance claims, mortgages, etc.) that today are usually held together with bespoke glue code and mismatched workflow tools, which are always about to be upgraded “any day now”.

Our platform replaces that glue code with a Common Operating Script, that runs reliably and consistently across teams and systems. Systems are connected to the platform rather than each other, so it’s easy to add and remove systems as your tools change, and changes must be approved by all teams before they go live, so everyone’s operating the same process, all the time.

To show how this works in production, we open-sourced a full Claims Settlement case study (75k loc in production, 12 systems and 5 teams), and you can check it out here: Repo: https://github.com/luthersystems/cross-department-claims-set... Video walkthrough: https://vimeo.com/1141432607?fl=pl&fe=cmt

How it works

• Logic: Write process-operations logic in a transactional Common Operating Script. Nodes execute this logic to process events across the various teams, while maintaining consistency.

• Infra: We provide a drop-in Kubernetes cluster with Prometheus/Grafana pre-wired. Use our managed version, or deploy to your cloud via Terraform.

• DevEx: Our connector hub provides 100s of ready-to-go connectors (S3, Postgres, etc.) and a prebuilt GitHub Actions pipeline. We provide a production-ready repo for you. We’d love feedback from the HN crowd, especially on the pain points you’ve seen when stitching together complex workflows in enterprise environments.