What AI-ready means
AI-ready describes the platform's ability to support search and analytics workloads that can use eligible Elastic AI and machine-learning features. It does not mean that every AI feature, model, or external model service is enabled automatically.
What the platform provides
Every deployment provides the Elasticsearch and Kibana foundation selected in the Marketplace plan. The deployment form currently distinguishes:
- a Search profile for full-text, vector, and AI-oriented search workloads;
- an Observability profile for logs, metrics, and traces;
- an optional Machine Learning setting.
The selected capacity, topology, Elastic version, license, and managed configuration determine what can be used after deployment.
Check feature availability
| Capability | Availability rule |
|---|---|
| Full-text search and analytics | Core Elasticsearch capability; confirm the deployed version |
| Vector fields and vector search | Confirm the deployed version and intended index design |
| Elastic machine learning | Requires the optional deployment setting and an eligible Elastic license |
| ELSER or other Elastic inference models | Use only when the deployed version, license, and model deployment are verified |
| External model providers | Requires customer-approved connectivity, credentials, data governance, and a supported integration |
| Kibana AI features | Confirm the deployed Kibana version, Elastic license, and managed configuration |
| AI-assisted Opsflw operations | Available only when explicitly exposed and approved for the managed application |
Do not infer feature availability from the product name or profile label. Check license and feature management and the compatibility reference before designing an AI workload.
Responsibility boundary
The customer owns the data, prompts, model-provider agreements, access policy, and use-case governance. iVedha operates the supported platform configuration. Elastic supplies the Elasticsearch and Kibana technology and defines which features belong to each Elastic subscription.
For a workload decision, record:
- the required Elastic capability;
- the deployment's Elastic version and license;
- whether ML or inference capacity is enabled;
- required outbound connectivity and credentials;
- data classification and model-provider approval;
- a representative performance and relevance test.
If any item is unknown, treat the capability as unavailable until it is verified.