Refactor usage instructions across deployments of a GenAIExample in the getting started guide.

[mkbhanda]

Whether we deploy a pipeline in a CSP or on-Prem, with one particular hardware or another, once the application is deployment, how we use it, test it, access each service endpoint is the same, presenting a refactor opportunity instead of duplicating that information per deployment option. Just highlight the difference between a Docker versus Kubernetes deployment in how the service endpoint is determined.