Bi Dashboard

BI Dashboard — Data Analytics Dashboard Full Pipeline

An agent team collaborates to generate data warehouse design, KPI definitions, visualizations, and automated reports in one pass.

Execution Mode

Agent Team — Five agents communicate directly via SendMessage and perform cross-validation.

Agent Composition

Workflow

Phase 1: Preparation (performed directly by the orchestrator)

1. Extract the following from user input:
   • Business domain: E-commerce/SaaS/Manufacturing/Finance/Education, etc.
   • Data sources: Available DB, API, file information
   • Key questions: Business questions the dashboard should answer
   • User tiers: Executives/team leads/practitioners — dashboard users
   • Existing assets (optional): Existing KPIs, data models, reports, etc.
2. Create the _workspace/ directory at the project root
3. Organize the input and save it to _workspace/00_input.md
4. If pre-existing files are available, copy them to _workspace/ and skip the corresponding phase
5. Determine the execution mode based on the scope of the request (see "Execution Modes by Request Scope" below)

Phase 2: Team Assembly and Execution

Tasks 3a (dashboard) and 3b (report) run in parallel.

Inter-agent communication flow:

• data-engineer completes > passes Measure/Dimension list to kpi-designer
• kpi-designer completes > passes KPI priorities/drill-down to dashboard-builder, passes reporting schedule/thresholds to report-automator
• dashboard-builder completes > passes snapshot capture method to report-automator
• bi-reviewer cross-validates all deliverables. On CRITICAL findings, requests corrections from the relevant agent > rework > re-verify (up to 2 rounds)

Phase 3: Integration and Final Deliverables

Finalize deliverables based on the reviewer's report:

1. Verify all files in _workspace/
2. Confirm that all CRITICAL findings have been addressed
3. Report the final summary to the user

Execution Modes by Request Scope

Reusing existing files: If the user provides existing KPI lists, data models, etc., copy them to the appropriate numbered location in _workspace/ and skip the corresponding agent.

Data Transfer Protocol

Error Handling

Test Scenarios

Normal Flow

Prompt: "Design a BI dashboard for an e-commerce store covering sales, customers, and inventory management" Expected result:

• Data model: E-commerce Star Schema (fact_orders, dim_product, dim_customer, dim_date)
• KPIs: Purchase funnel (visits > cart > checkout), revenue, average order value, repurchase rate, inventory turnover
• Dashboard: Executive Summary + Sales + Customer + Inventory, 4 tabs
• Reports: Daily sales summary, weekly team report, monthly executive report
• Review: Full consistency matrix verification

Existing File Reuse Flow

Prompt: "Design a dashboard layout with this KPI list" + KPI file attached Expected result:

• Copy existing KPIs to _workspace/02_kpi_definition.md
• Visualization mode: Deploy dashboard-builder + reviewer
• Skip data-engineer and kpi-designer

Error Flow

Prompt: "Build a dashboard for our company" Expected result:

• Domain/data source unknown > propose 3 common business domains, guide selection
• After selection, proceed with full pipeline based on domain-standard schema

Agent Extension Skills