Home › Key concepts › One framework, many customers
Key concept · Run (advanced) · the same on every platform
One framework, many customers
Build reusable pieces once; bind them to each customer's cloud and rules.
What it is
Every engagement repeats the same needs: ingestion, evaluation, guardrails, tracing. Keep these as a framework with ports, and write a small adapter and configuration per customer.
Why it matters
It makes the second customer faster and cheaper than the first, and improvements reach everyone.
An example
The evaluation harness built for one insurer runs at a bank with a new rules file and a different trace adapter.
The common mistake
Forking the code for each customer.
Where you meet it on each platform
The idea stays the same; only the product name changes.
| Job | AWS | Azure | Google Cloud | Databricks | Open source |
|---|---|---|---|---|---|
| Infrastructure as codethe whole setup in files you can review and re-create | Terraform or AWS CDK | Bicep or Terraform | Terraform | Declarative Automation Bundles (Asset Bundles) | Docker Compose |
| Agent runtimewhere a long-running agent keeps its state and waits safely | Bedrock AgentCore | Foundry Agent Service / Durable Functions | Vertex AI Agent Engine | Mosaic AI Agent Framework | DBOS or Temporal + your loop |
Related ideas
Where it shows up
Part of the Run level on these paths: Forward Deployed Engineering (scale).