System Datasources
Beak AI Studio enables agents to connect with enterprise platforms and business systems, allowing seamless integration with core business processes and data.
Datasource Creation Workflow
Creating a System datasource in AI Studio involves two main phases to ensure robust integration and knowledge management.
Phase 1: Initiation
Start by providing the following essential information:
- Datasource Name: A clear, descriptive name for the system integration.
- Description: A summary of the datasource’s purpose and intended use.
- Usage Instructions: (Optional) Guidance for team members on how and when to use this datasource.
- Datasource Type: Select "System" as the type.
Once this information is entered, proceed to the configuration phase.
Phase 2: Configuration
A. System Connection
- Connection Parameters: Enter required connection details (API endpoints, authentication, etc.) for the enterprise system.
- Supported Integration Methods:
- REST API: Connect to systems exposing RESTful endpoints.
- OpenAPI/Swagger: Import OpenAPI specifications for rapid integration.
- Custom API: Configure custom endpoints and authentication as needed.
- Authentication Credentials:
- API Key, OAuth2 client ID/secret, or username/password, depending on the platform.
- Permissions: Ensure the API user has sufficient permissions for required operations (read, write, update, etc.).
- Network Access: The Beak AI Studio environment must be able to reach the system's API endpoint (consider VPN, firewall, or IP allowlisting if needed).
- Validation: The system tests the connection and validates credentials and permissions.
B. Resource Selection (Subsequent Step)
After the System datasource is established and validated, choose which collection of records to ingest:
- Resource: Pick the collection (tickets, devices, users, and so on). Connectors that declare their resources offer them as a list; otherwise enter the resource name.
- Detect automatically: Beak AI can probe the API and report what is available — where the records are, which fields exist, how it paginates, and which field marks a record as changed. See Custom Connectors.
- Validation: A datasource with no resource selected cannot sync, and says so.
A resource must be saved before the first sync.
The System datasource becomes available for use only after both the connection and resource selection steps are successfully completed and validated.
Supported System Integrations
Beak AI Studio supports integration with the following platforms. For each, the typical integration method is via REST API or OpenAPI, with authentication and data types as described.
| Platform | Integration Method(s) | Supported Data Types | Configuration Steps |
|---|---|---|---|
| Atlassian Jira | REST API, OpenAPI | Issues, projects, users, comments | API endpoint, OAuth2/API token, permissions |
| Azure | REST API, OpenAPI | Resources, metrics, logs | API endpoint, OAuth2, Azure AD, permissions |
| Azure Log Analytics | REST API | Logs, queries, metrics | Workspace ID, API key, endpoint, permissions |
| Azure Metrics | REST API | Metrics, time series | Resource ID, API key, endpoint, permissions |
| Beak FinOps | REST API | Cost, usage, billing data | API endpoint, API key, permissions |
| Beak ITSM | REST API | Tickets, incidents, assets | API endpoint, API key, permissions |
| Beak Observability | REST API | Metrics, traces, logs | API endpoint, API key, permissions |
| Beak RMM | REST API | Device status, alerts, inventory | API endpoint, API key, permissions |
| Connectwise Manage | REST API, OpenAPI | Tickets, contacts, agreements | API endpoint, API key, company ID, permissions |
| ManageEngine Endpoint Central | REST API | Devices, patches, compliance | API endpoint, API key, permissions |
| ManageEngine ServiceDesk Plus | REST API | Tickets, requests, assets | API endpoint, API key, permissions |
| Microsoft 365 | REST API, OpenAPI | Users, mail, calendar, files | API endpoint, OAuth2, Azure AD, permissions |
| Microsoft Entra | REST API | Identities, access policies | API endpoint, OAuth2, Azure AD, permissions |
| Microsoft Intune | REST API | Devices, compliance, apps | API endpoint, OAuth2, Azure AD, permissions |
| N-Central | REST API | Devices, monitoring, alerts | API endpoint, API key, permissions |
| Ninjaone | REST API | Devices, tickets, monitoring | API endpoint, API key, permissions |
| Plumsail | REST API | Forms, workflows, documents | API endpoint, API key, permissions |
| Qualys | REST API | Vulnerabilities, assets, scans | API endpoint, API key, permissions |
| SentinelOne | REST API | Endpoints, threats, policies | API endpoint, API key, permissions |
| ServiceNow | REST API, OpenAPI | Tickets, incidents, CMDB, users | API endpoint, OAuth2/API key, permissions |
| SolarWinds Service Desk | REST API | Tickets, assets, users | API endpoint, API key, permissions |
| Webhook | Webhook URL | JSON payloads | Webhook URL, authentication (if required) |
| Windows Defender | REST API | Alerts, device status, threats | API endpoint, API key, permissions |
| Zendesk | REST API, OpenAPI | Tickets, users, organizations | API endpoint, OAuth2/API token, permissions |
| Zoho Desk | REST API | Tickets, contacts, knowledge base | API endpoint, API key, permissions |
Other External Systems:
Any other external system can be integrated without waiting for a product release — register a connector manifest and it appears in the connector picker straight away. Beak AI can auto-detect what the application exposes rather than requiring a hand-written data mapping. See Custom Connectors.
Platforms That Sync Records Today
Connecting to a platform and ingesting its records are two separate capabilities. Every platform above accepts a connection and validates it. The platforms below additionally declare a resource, so they can sync records into the vector store for agents to retrieve:
| Platform | Category | Resource | Ingested as |
|---|---|---|---|
| Atlassian Jira | ITSM | Issues | ticket |
| Beak ITSM | ITSM | Tickets | ticket |
| ConnectWise Manage | ITSM | Service Tickets | ticket |
| ManageEngine ServiceDesk Plus | ITSM | Requests | ticket |
| ServiceNow | ITSM | Incidents | ticket |
| SolarWinds Service Desk | ITSM | Incidents | ticket |
| Zendesk | ITSM | Tickets | ticket |
| Zoho Desk | ITSM | Tickets | ticket |
| N-Central | Asset Management | Devices | asset |
| SentinelOne | Security | Agents | asset |
A platform not on this list still connects — it just has no resource to ingest yet, so Sync now has nothing to fetch. Use Detect automatically to see what its API exposes, and Custom Connectors to declare a resource for it without waiting for a release.
Sync Behaviour
Every system sync is a complete refresh: it re-reads the whole resource, then replaces the datasource's previous records with what it just read. One copy of every record, matching the source system as of that run — edits are corrected, deletions disappear, and running twice does not double the count.
Reading only what changed since the last run is not offered yet. Six of the ten resources above expose a last-changed field and four do not, and for at least one platform it is unconfirmed whether the vendor's date filter covers edits or only record creation — a filter covering only creation would silently never re-read an edited ticket. Until that is settled per vendor, one correct mode beats two where one is quietly lossy.
Full detail, including what happens when a safety limit cuts a run short: Custom Connectors — Sync.
Example Configuration Steps
-
Select System Platform:
Choose the desired platform from the supported list. The picker groups platforms under category headings — ITSM, Asset Management, Knowledge Base, Security, Identity, DevOps, Database, Other — so a platform is found by the job it does rather than by scanning one long alphabetical list. A connector with no category set appears under Other. -
Enter Connection Details:
Provide the API endpoint, authentication credentials (API key, OAuth2, etc.), and any required parameters. -
Set Permissions:
Ensure the API user or app registration has the necessary permissions for the required operations. -
Select the Resource:
Choose which collection to ingest, or use Detect automatically to have Beak AI report what the API exposes. -
Save and Enable:
Save the datasource. It is now available for use in agent workflows.
Best Practices
- Use OAuth2 or API keys for secure authentication.
- Limit API user permissions to only what is required.
- Regularly rotate credentials and monitor access logs.
- Use Detect automatically to confirm the record path and incremental field rather than entering them by hand.
- Size the sync schedule to the resource — every run re-reads the whole collection, so a large one on a frequent schedule is real load on the source system's API limits.