Automated Inbox Triage Agent

The Intelligent Communications Triage Agent is an automated, AI-driven pipeline designed to solve the critical "signal-to-noise" problem in high-volume communication channels. Rather than relying on rigid keyword filters, this system leverages a Large Language Model (LLM) to perform semantic analysis on inbound messages, categorize user intent, and dynamically route the data based on priority. It ensures instant execution for critical alerts while silently aggregating routine data into a structured database, effectively eliminating alert fatigue and reducing mean-time-to-resolution (MTTR).

System Architecture & Data FlowThis multi-tool automation is orchestrated to process unstructured text and convert it into highly actionable, structured data streams:Ingestion Layer (Gmail API): Monitors inbound traffic using strict search parameters to isolate relevant telemetry, reducing unnecessary API calls and processing overhead.Cognitive Processing (OpenAI API): Unstructured email payloads are passed to an LLM prompted to act as a triage engineer. The AI enforces a strict JSON schema output, extracting specific entities: Intent Category, Sender Entity, Actionability (Boolean), a 1-sentence TL;DR summary, and the primary action URL.Dynamic Routing Logic: The system evaluates the AI's structured output against defined conditional logic to determine the execution path.Path A (High-Priority/Critical): For high-value alerts (e.g., direct recruiter outreach or application updates), the system immediately pushes a formatted payload via the WhatsApp Business API, delivering the TL;DR and action link directly to a mobile device.Path B (Data Aggregation): For routine alerts (e.g., automated job recommendations), the system bypasses the notification engine to prevent alert fatigue. Instead, it writes the extracted data directly to a Google Sheets database.Path C (Network Noise): Non-actionable updates are ignored entirely.Strategic Impact & Downstream IntegrationThis architecture demonstrates advanced automation capabilities by moving beyond linear "trigger-action" setups into intelligent, context-aware routing. By converting messy inbox data into a clean, localized database, the system lays the groundwork for modular expansion. The Google Sheet tracking routine job alerts now serves as an automated ingestion layer for broader workflows, acting as the perfect data feed for an AI Job Application Copilot to subsequently pull those listings and score them against a resume.