Create MCP-powered JobPilot analytics bot
Goal
Create an AI bot using MCP that can answer questions about JobPilot data and retrieve the required information through structured queries.
Example questions
- How many applications did I submit this month?
- How many companies am I currently interviewing with?
- What is my interview rate?
- Which companies rejected me?
- How many interviews did I have last month?
- Which roles had the highest fit score?
- How do referrals affect my interview rate?
- How many applications have been inactive for more than X days?
Requirements
The AI should:
- Understand the user's question.
- Determine what data is required.
- Generate/use the appropriate MCP query/tool.
- Retrieve the data.
- Validate the result.
- Produce a concise answer.
- Avoid inventing data when the database does not contain enough information.
Security
- MCP tools must enforce user-level data isolation.
- The bot must never access another user's data.
- Query inputs must be validated.
- Avoid arbitrary database access where possible.
- Prefer predefined, safe query tools.
Acceptance Criteria
- Bot can answer common analytics questions.
- Queries retrieve actual JobPilot data.
- No hallucinated statistics when data is unavailable.
- User isolation is enforced.
- MCP calls are logged/auditable where appropriate.
- Tests cover representative questions.
Create MCP-powered JobPilot analytics bot
Goal
Create an AI bot using MCP that can answer questions about JobPilot data and retrieve the required information through structured queries.
Example questions
Requirements
The AI should:
Security
Acceptance Criteria