OpsDoctor: Your AI Operations Doctor
Link to open source: https://github.com/shubhamranswal/opsdoctor
OpsDoctor is an agentic AI operations assistant that investigates business issues, reasons across multiple data sources, and recommends the next best action with evidence.
Instead of simply displaying failed payments or generating static reports, OpsDoctor autonomously investigates operational signals, identifies affected customers, retrieves relevant customer and business context, checks the Knowledge Base and policies, and produces a validated recommendation for the operations team.
For our demo, OpsDoctor focuses on payment failures and customer retention. When payment failures occur, the agent investigates the affected transactions, gathers supporting customer context, evaluates potential churn risk, retrieves the relevant retention policy, and validates the proposed action before requesting human approval.
The agent can work across integrations such as payment providers, Gmail, Slack, Jira, and Notion through Swytchcode. Its reasoning loop follows an agentic workflow: understand the problem, select the appropriate tools, execute them, analyze the results, gather additional evidence when required, validate the proposed action, and then take the next step.
The goal is to move from "What went wrong?" to "What should we do next, and why?" while keeping humans in the loop for consequential actions.
Built with an agentic architecture using Swytchcode integrations, OpsDoctor demonstrates how AI can turn fragmented operational data into actionable, evidence-backed decisions.
