Status
Live
Production-ready
Data & Analytics
Report Quality Agent
Final quality gate that validates narrative content, filters technical terms from all user-facing text, ensures minimum insight counts, validates recommendations, and scores overall report quality on a 0-to-100 scale. No
Enterprise AssistantLiveInternal (Colaberry Enterprise)Verified
Department
Data & Analytics
Enterprise Assistant department for Colaberry Enterprise agents
Source
Internal (Colaberry Enterprise)
Built by Colaberry
About
About the Agent
What this agent does, the challenges it addresses, and where it delivers value.
Final quality gate that validates narrative content, filters technical terms from all user-facing text, ensures minimum insight counts, validates recommendations, and scores overall report quality on a 0-to-100 scale. No LLM calls - pure functions with sub-2ms latency.
Challenges This Agent Addresses
- 1**Executive Reporting**: Ensures C-suite dashboards never display technical jargon like "NULL" or "PostgreSQL"
- 2**Business Intelligence**: Guarantees every report has actionable insights even when the underlying queries return minimal data
- 3**Quality Monitoring**: The quality score enables tracking report quality over time and flagging degradation
Workflow
How the Agent Works
Step-by-step operational flow showing how this agent processes tasks end-to-end.
1
Step 1
**Technical term filtering**: Applies 20+ regex-based replacements to remove or translate technical terms:
2
Step 2
Database terms (SQL, Sequelize, PostgreSQL, schema) become neutral alternatives
3
Step 3
Programming terms (null, undefined, NaN, bigint) become "no data"
4
Step 4
Agent internals (agent_name, trace_id, stack_trace) become business terms
5
Step 5
Cleans up artifacts like double spaces, double commas, and leading punctuation
6
Step 6
**Narrative section filtering**: Applies the same technical term filter to executive summary, detailed analysis, key findings, risk assessment, recommended actions, and follow-up areas
7
Step 7
**Minimum insight guarantee**: If fewer than 3 insights exist, synthesizes additional ones from SQL results:
8
Step 8
Single-row aggregates become KPI-style insights ("Total Leads: 1,234")
9
Step 9
Multi-row results become highlight insights ("Top Agent: CoryBot with 500 Executions")
10
Step 10
**Recommendation validation**: If no recommendations exist, provides sensible defaults. Filters technical terms from existing recommendations and removes any that become too short after filtering.
11
Step 11
**Quality scoring** (0 to 100): Awards points for narrative length (20), chart count (20), insight count (20), technical term absence (20), and recommendation count (20)
Execution Modes
Trigger: event (invoked as the last step in query response assembly)
Data
Inputs & Outputs
What data this agent consumes and the artifacts or actions it produces.
Input Data
- `AssistantResponse` object containing narrative, narrative sections, insights, visualizations, and recommendations
- Optional `SqlResult` list for synthesizing additional insights if needed
Deliverables
- Validated `AssistantResponse` with:
- Technical terms filtered from all narrative text
- Minimum 3 insights guaranteed (synthesized from SQL if needed)
- Recommendations validated for business relevance
- All insight messages cleaned of technical language
Core Tasks
- Data & Analytics
Integrations
Systems Connected
Internal systems, APIs, and tools this agent integrates with.
Tools & APIs
Invoked by the **Query Engine** as the final validation stepUses **Data Analyst Agent** (`getBusinessLabel`) for insight enrichmentProcesses output from **Chart Validation Agent** (visualizations are already validated)Quality score can be logged by **Process Observation Agent** for monitoring
Specifications
Agent Specs
Technical specifications, requirements, and deployment details.
Related
Related Agents
Other agents in the same department or industry.
Enterprise AI
Ready to deploy this agent?
Schedule a walkthrough with our team to see how this agent integrates with your workflows.