Data & Insights

Data & Insights is the Slab5 area for turning application events and workspace records into operational analytics. It gives builders a path from raw product activity to Object Analytics, tracked objects, reusable BI views, scheduled insights, Analytics Sites, and Custom BI templates.

Use it when you want to answer questions such as:

• Which objects in my app are getting the most activity?
• Which landing pages, forms, accounts, tickets, projects, or workflow runs are active?
• Which events are happening by app, object type, object ID, and time window?
• Which operational signal should become a reusable BI view?
• Which insights should be refreshed on a schedule?
• Which governed metric answers a natural-language business question?

The module is workspace-scoped. Events, insight views, BI objects, schedules, and analytics sites stay inside the workspace selected by the user or API credential.

Console areas

The Slab5 console currently organizes Data & Insights into these main areas:

Inside Insights, users can move between Object Analytics, Insight Builder, and Insight Views. Analytics Sites is promoted to its own Data & Insights console area because it manages event-collection site configuration, allowed origins, and site-level access.

Data & Insights is separate from the Data Agents API module. Data & Insights is the console analytics experience; Data Agents API is the developer and agent-facing module for governed datasets, lineage, execution jobs, decisions, recommendations, and knowledge artifacts.

Explore

Explore is the natural-language BI surface in Data & Insights. It uses governed Slab5-native metric definitions, semantic retrieval, deterministic SQL compilation, managed query execution, and saved answers.

Explore is intentionally metric-level only. It indexes approved metric definitions, SQL expressions, dataset descriptions, schema descriptions, glossary terms, and example questions. It does not embed raw BI rows or copy raw lake files into Slab5.

Use Explore for questions such as:

• How many records did we have last quarter?
• What was our deal count this month?
• How many support tickets were opened last week?
• What project count changed over the last 30 days?

Explore readiness

Before a user asks a question, the Readiness tab checks:

• governed metrics exist
• the Slab5 Vectors semantic index is active
• a Slab5 Query Engine execution plane is active
• the internal API bridge is configured
• the workspace is using Slab5-native BI datasets for this release

Use readiness before production use, after adding new metrics, or after deploying a new environment.

Nine-stage pipeline

Each Explore run follows the BI Intelligence pipeline:

The Pipeline runs tab shows persisted BI Explore execution jobs. Open View trace to inspect stages, status, parsed plan, logs, rows, and lineage after a page refresh.

Saved answers

Completed Explore runs are saved as BI Intelligence answers. Saved answers include:

• the original question
• selected metric and dataset references
• result rows when available
• narration
• pipeline stages
• deterministic SQL lineage
• source references

Use Saved answers when a user wants to revisit a prior answer without rerunning the question.

Managed execution

Explore uses managed Data Agent execution for the query stage. In production, the API creates a BI execution job and dispatches a Slab5 Query Engine worker task. If the worker succeeds, the pipeline reaches narration and insight persistence. If it fails, admins should inspect the Pipeline runs trace and the Data Agent execution logs.

Object Analytics

Object Analytics is a generic application-event explorer. It is not tied to a single app, CRM, marketing site, or workflow product. Developers can emit events from any product surface and use metadata to describe what the event belongs to.

An event can identify:

• app: a developer-defined application or surface, such as marketing-site, crm, customer-portal, support-console, or workflow-app.
• event_name: the action that happened, such as page.viewed, button.clicked, form.submitted, lead.created, ticket.opened, or workflow.started.
• object_type: the thing being acted on, such as landing_page, form, lead, account, ticket, workflow, or campaign.
• object_id: the stable object identifier in the builder's app.
• properties: additional event details.

The Object Analytics table groups events by object so users can see which tracked objects are active over the selected window. It is already aggregated data, not a raw event stream. The grouping key is effectively:

app + object_type + object_id

That lets a builder compare activity across pages, forms, customers, workflows, or any object model they define.

Click to enlarge

Event collection

Applications send events into Slab5 through the application-event collection surface. A typical event payload includes the app, event name, object context, user or session context, and arbitrary properties.

curl -X POST "$SLAB5_API_BASE_URL/v1/insights/app-events" \
  -H "Authorization: Bearer $SLAB5_API_KEY" \
  -H "Content-Type: application/json" \
  -H "x-slab5-workspace-id: $SLAB5_WORKSPACE_ID" \
  -d '{
    "app": "marketing-site",
    "event_name": "form.submitted",
    "object_type": "form",
    "object_id": "beta-signup",
    "properties": {
      "source": "homepage",
      "plan_interest": "growth"
    }
  }'

Use consistent event names and object identifiers. Slab5 does not require developers to register a formal app object before sending events; the app value is a developer-defined grouping key used for filtering and reporting.

Tracked objects

The tracked objects view is designed for quick inspection:

• Filter by app to focus on one product surface.
• Filter by object type to compare similar entities, such as forms or landing pages.
• Filter by object ID when investigating one object.
• Change the time window to compare recent activity.
• Use pagination when the workspace has many tracked objects.

Examples:

Promote to BI

Promote to BI turns an Object Analytics view into a saved insight view. This is useful when a user discovers a useful operational slice and wants it to become reusable instead of re-entering filters every time.

The recommended flow is:

1. Open Data & Insights → Insights → Object Analytics.
2. Set the app, object type, object ID, and time window filters.
3. Review the tracked objects table.
4. Click Promote to BI.
5. Name the insight view and confirm the source filters.
6. Open the promoted view in Insight Builder or Insight Views.
7. Schedule refresh if the view should update automatically.

Promoted views start as operational insight definitions. They can then be refreshed and scheduled like other BI-backed views.

Click to enlarge

Insight Builder

Insight Builder is where users inspect or configure saved insight definitions. It is the bridge between an operational signal and a reusable analytics view.

Use Insight Builder to:

• Review the source of an insight view.
• Confirm whether the view came from Object Analytics or another dataset.
• Inspect the configuration that defines the view.
• Open related saved views and scheduled refresh settings.

Click to enlarge

Insight Views

Insight Views are saved analytics definitions that can be reused across a workspace. A view may come from Object Analytics, a Custom BI template, or another supported dataset source.

Each view should make it clear:

• What question it answers.
• Which source data it uses.
• Whether it has refreshed yet.
• When it last refreshed.
• Whether it is scheduled.
• Whether the most recent refresh succeeded or failed.

Scheduling

Insight views and BI artifacts can be refreshed on a schedule. Current scheduling options include:

• Hourly
• Daily
• Weekly
• Monthly

For monthly schedules, use Last day of month when the schedule should run at month end regardless of whether the month has 28, 29, 30, or 31 days.

Click to enlarge

Analytics Sites

Analytics Sites are surfaces for packaging analytics into a site-like experience. They are useful when a team wants to share a curated analytics view with a specific audience instead of sending users into the full console.

Use Analytics Sites for:

• Customer-facing reporting portals.
• Internal executive dashboards.
• Campaign reporting pages.
• Client reporting for agencies.
• Embedded analytics experiences in vertical apps.

Click to enlarge

Custom BI

Custom BI uses Slab5 operational data for workspace analytics. Built-in templates are launchable without external connectors because they read workspace-scoped Slab5 records.

Current Slab5-native templates include:

• content_performance: CMS entries, marketing posts, and campaigns by status, channel, content type, published count, and budget.
• crm_pipeline: deals by stage, stage key, currency, count, and value.
• support_ticket_health: tickets by status and priority with first-response and resolution SLA risk counts.
• usage_activation: usage events by date, event type, route, method, credential path, units, and billed usage.
• agentops_execution_health: Data Agent execution jobs by job type, status, trigger, attempts, and average duration.

Click to enlarge

Click to enlarge

Custom BI runs

Runs show the execution history behind Custom BI refreshes, scheduled jobs, exports, and Data Agent-backed BI work. Use this view to confirm whether a refresh was queued, running, completed, or failed, and to inspect attempts, trigger source, queued time, finish time, related objects, and job details.

Custom BI runs are useful when a team needs to answer:

• Did this report or metric refresh successfully?
• Was the job triggered manually, by schedule, or by an analytics pipeline?
• How many attempts did the job need?
• When was the job queued and when did it finish?
• Which related asset, view, or pipeline produced the run?

Click to enlarge

REST resources

Use bi:read to list BI objects and bi:write to create, update, archive, run queries, create exports, and manage scheduled artifacts.

Common BI endpoints:

• GET /v1/bi/dashboards
• POST /v1/bi/dashboards
• GET /v1/bi/datasets
• POST /v1/bi/datasets
• GET /v1/bi/reports
• POST /v1/bi/reports
• GET /v1/bi/metrics
• POST /v1/bi/metrics
• GET /v1/bi/queries
• POST /v1/bi/queries/run
• GET /v1/bi/exports
• POST /v1/bi/exports
• GET /v1/bi/scheduled-artifacts
• POST /v1/bi/scheduled-artifacts
• PATCH /v1/bi/{collection}/{id}
• DELETE /v1/bi/{collection}/{id}

Application event and Object Analytics endpoints:

• POST /v1/insights/app-events
• GET /v1/insights/app-events

The Object Analytics read endpoint returns grouped object analytics for the selected app, object filters, and time window.

MCP tools

The BI module is also available to MCP clients through these tools:

• list_bi_dashboards
• create_bi_dashboard
• list_bi_reports
• create_bi_report
• list_bi_datasets
• create_bi_dataset
• list_bi_metrics
• create_bi_metric
• run_bi_query
• create_bi_export
• update_bi_object
• archive_bi_object

Examples

Run a built-in workspace dataset query:

curl -X POST "$SLAB5_API_BASE_URL/v1/bi/queries/run" \
  -H "Authorization: Bearer $SLAB5_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "dataset_key": "api_usage",
    "limit": 25
  }'

Run a Slab5-native Custom BI template:

curl -X POST "$SLAB5_API_BASE_URL/v1/bi/queries/run" \
  -H "Authorization: Bearer $SLAB5_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "dataset_key": "content_performance",
    "limit": 25
  }'

Create a BI export artifact:

curl -X POST "$SLAB5_API_BASE_URL/v1/bi/exports" \
  -H "Authorization: Bearer $SLAB5_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Weekly activation CSV",
    "format": "csv",
    "query_id": "query_123"
  }'

Traceability

Slab5 records Data & Insights activity with workspace context and request IDs. This helps teams trace event collection, BI query runs, scheduled refreshes, MCP tool calls, and API requests when reviewing usage, debugging integrations, or supporting enterprise audit workflows.