Connectors & Third-Party Data Source Integration

**Value proposition**
Agent Builder is expanding beyond "chat with your Elasticsearch data" to "chat with your entire ecosystem." By connecting agents to tools like Slack, Gmail, HubSpot, Confluence, and more, users can ask questions and take actions across the tools they use every day — without leaving Elastic. These connectors use a federated approach, meaning they query data with the supplied credentials through the Agent when requested. This allows the data to stay in the source and respect all appropriate security and access controls.
Agent Builder leverages its agentic search capabilities to optimize results from across these sources.

**Expected outcome**
A growing library of first-class connectors that agents can use through Workflows. Connectors are surfaced natively in Agent Builder, support per-user OAuth authentication, and can be dynamically invoked in chat via `@connector` mentions.

Planned Connectors (list to be expanded):
- Slack
- Teams
- Google Drive
- Gmail
- Jira Cloud
- Salesforce CRM
- ServiceNow
- Sharepoint Online
- Web Search (Tavily)
- Web Search (Jina Reader)
- SharePoint Server
- Pager Duty
- Hubspot
- Notion
- Azure Blob Storage
- Google Cloud Storage

**Key user stories**
- As a sales rep, I want to ask "Which of my high value customers that have declining usage in the last 30 days?" and get an answer from across Salesforce & Elasticsearch data
- As a developer, I want to say "Post my standup update to Slack based on my GitHub PRs today" and have thee agent handle it end-to-end.
- As an admin, I want users to authenticate to third-party services with their own credentials so data access is properly scoped.