HOW LUMI ANSWERS QUESTIONS
Lumi uses AI agents to interpret user intent, translate plain language prompts into code, and deliver insights.
User Query: The user asks a question in plain language.
Intent Understanding: Lumi’s AI ensures a thorough understanding of the user's requirements.
Context Retrieval: The system retrieves relevant context from the knowledge base.
Code Generation: Lumi generates the necessary code to address the query.
Execution and Results: The code is executed within the client's network, and the results are processed.
Insight Delivery: Lumi summarizes and delivers key findings and insights back to the user.
KNOWLEDGE BASE
The knowledge base is the map that guides Lumi through your data. Without it, automated responses to data questions are not possible.
The knowledge base contains information about:
Data-model Schemas
Field Definitions
Logic to Calculate Business KPIs
Specific Business Terminology
NOW AVAILABLE
Managed knowledge base services available to enterprise customers.
Bonus
Lumi's knowledge base is enriched with domain expertise from experienced data scientists and industry experts.
core technology: ai agents
Deciphering True Intent & Delivering Reliable Insights
Our IP lies in translating user desires into precise, consistent, and reliable AI-driven results.
Lumi consists of a network of AI agents "agentic workflows" that work together to understand the true intent of the user and guarantee accurate, reliable, and repeatable insights.
Lumi has been designed to mimic the characteristics of a highly capable human analyst.
Lumi seeks clarifications and suggests alternatives when vague requirements are given.
Lumi summarizes the key findings from the output - “the so what?”
Lumi provides complete visibility into the assumptions and logic used to derive the insights.
Meet some of the AI agents Lumi employs to ensure accurate, reliable, and actionable insights.
Here are some (not all) of our agents for your references:
Ensures understanding of user requirements
Retrieves relevant context from the knowledge base.
Generates the necessary code to answer the question.
Summarizes key findings and insights for the user.
