Business Rules & Metadata
Brand team–defined rules and dataset metadata that guide how the agent interprets and processes queries to ensure accuracy and consistency.
ENGINE
LAILA for Sub-National Sales Analytics
App Level
App Level
Engine involves multi-agent architecture that handles multiple tasks and uses dynamic routing (supervisor agent)
Capability Level
Engine only involves 1 or 2 agents in sequence and does not require routing
App Level
Engine involves multi-agent architecture that handles multiple tasks and uses dynamic routing (supervisor agent)
Capability Level
Engine only involves 1 or 2 agents in sequence and does not require routing
Get fast, accurate answers to your subnational analytics questions using natural language.
This engine empowers analytics users to easily interact with Sub-National datasets through natural language instead of SQL. Each question is answered with an insightful summary, a downloadable data table, and visualizations to highlight key patterns, comparisons, and trends. it also provides the auto-generated SQL query and a clear explanation for transparency. Users can perform multi-step analytics with context retention, retained, enabling seamless follow-up questions to further refine or expand insights.This engine simplifies access to complex structured data, speeds up decision-making, and reduces dependency on technical teams.
Engine Prerequisites
Metadata and any other required context or configurations that must be onboarded to use this engine
Dataset Knowledge
Embedded understanding of structured tables, enabling precise mapping of natural language questions to governed metrics.
Compliance Boundaries
Strict guardrails to prevent out-of-scope or non-compliant queries, ensuring responses remain valid and aligned with approved data usage.
Structured Data Access
Direct connection to curated, schema-driven datasets.
Workflow
A specific example of how to this engine processes a prompt, selects and configures the appropriate agents and tools, and utilizes them to achieve the desired output.
Legend
Prompt
Agent
Supervisor Agent
Tools
Output
Inputs & Outputs
Example Prompts
What is the CLL National VEN TRx Share?
Results or Output
Type of Output
Text Summary
Description
AI-generated clear and insightful summary of reponse data
Type of Output
Data Table
Description
Data gathered from the generated SQL query
Type of Output
Data Visualization
Description
Graph of data if applicable
Type of Output
SQL Query
Description
Generated SQL query
Type of Output
Query Explanation
Description
Simple, step-by-step explanation of generated SQL logic
