Agentic Radar
AI-powered Executive Dashboard for 24/7 supply chain risk mitigation.
Intelligent early-warning supply chain risk engine based on the international IEC 62402 standard. Ingests End-of-Life notices (PDN/EOL) via webhooks and emails, extracts critical MPNs using CrewAI cognitive agents, traverses the PostgreSQL BOM graph, and computes financial revenue at risk in real time.
System Objectives
Eradicate the 'Black Swan' fallacy in manufacturing operations by quantifying technical obsolescence as predictable entropy under IEC 62402.
Ingest multichannel alerts across both structured sources (commercial B2B APIs) and unstructured raw manufacturer emails.
Traverse the BOM relational graph to map obsolete semiconductor MPNs directly to impacted revenue-generating end products.
Translate autonomous agent telemetry into actionable financial P&L metrics for C-level executives (CFO, COO, VP Supply Chain).
Key Capabilities & Modules
Event-Driven Architecture (EDA): asynchronous webhook endpoints in FastAPI preventing external system timeouts.
CrewAI Cognitive Agent: semantic extraction of target part numbers (MPNs) from unstructured text using LLMs.
Margin-at-Risk Computation: instant computation of threatened financial margins (e.g. 15,450 EUR) mapped to assemblies.
Assembly Blacklist: classification of operational impact (CRITICAL LTB REQUIRED vs EVALUATING INVENTORY).
Live Telemetry Feed: animated sequential terminal displaying cognitive agent reasoning steps and SQL tool calls.
Deterministic 24/7 Sandbox Mode: lightweight tactical UI in Tailwind and Vanilla JS loading instantaneously on static CDN.
Architecture & Design Specifications
Decoupled 4-tier architecture: asynchronous ingestion layer with FastAPI receiving JSON webhooks and inbound emails, CrewAI cognitive orchestrator executing background workers with SQL graph tools, relational PostgreSQL in Supabase modeling multilevel BOM trees (schema obs_radar), and tactical executive dashboard served on static CDN with deterministic sandbox fallback.
🔍 Click diagram to enlarge to fullscreen
Invariant Architectural Principles:
- Strict schema isolation: component and BOM tables reside exclusively within the dedicated 'obs_radar' schema.
- Deterministic financial computing: margin-at-risk aggregation is computed in precise database queries, not LLM estimation.
- Continuous availability (Dual-Mode): the public web app never displays 'OFFLINE' states, operating with deterministic synthetic data.
Technology Stack
FRONTEND
BACKEND
DATA
AI
Reference Articles & Architecture Series
This application is fully documented in an engineering series published on the blog:
The Chess of Obsolescence: Turning Supply Chain Collapse into Your Ultimate Competitive Advantage
The narrative dominating the corridors of large modern corporations is dangerously flawed. When a production line, manufacturing …
Read full article →
The Tactical Arsenal: Why Buying Supply Chain Radar Tools Won't Save Your Production
1. The Hook: The False Sense of Security 📡 Official Datalaria Application Agentic Radar v1.0.1 AI-powered Executive Dashboard for …
Read full article →
The Risk Map: Architecting an Obsolescence-Immune Data Foundation
1. The Flat Table Trap (Excel is Dead) 📡 Official Datalaria Application Agentic Radar v1.0.1 AI-powered Executive Dashboard for …
Read full article →
Trial by Fire: From Garbage Excel to Relational Graph with Python and Pandas
1. The Hook: Industrial Data Entropy 📡 Official Datalaria Application Agentic Radar v1.0.1 AI-powered Executive Dashboard for 24/7 …
Read full article →
The Agentic Radar: Why LLMs Won't Save Your Supply Chain (And Tool Calling Will)
Destroying the ‘Chat with PDF’ myth: How to orchestrate Gemini 2.5 via CrewAI so it delegates deep mathematical responsibilities …
Read full article →
The Central Nervous System: Scaling the Agentic Radar to 24/7 with FastAPI and Webhooks
Transforming the Obsolescence radar into an Event-Driven 24/7 unified microservice. We cover dual ingestion vectors (Accuris vs. …
Read full article →
The Executive Interface: Translating AI Telemetry into C-Level Decisions
The final block in Obsolescence Engineering: Projecting the mathematical engine of the agentic radar towards a lightweight, …
Read full article →Evolutionary History & Versions (SemVer)
Formal record of releases, feature additions, and security patches:
SDD isolation, static/ firewall and 24/7 deterministic Sandbox mode
- Backend code isolation in apps/obs-management/api/ eradicating .py, .sql and .env files from static/.
- Permanent eradication of the OFFLINE connection error by enabling client-side deterministic Sandbox mode.
- Navigation integration with Datalaria catalog and contextual upsell banners to B2B executive packs.
Official public release backed by the 7-article engineering series
- Development of the CrewAI agentic orchestration engine and semantic MPN extraction.
- Relational BOM tree modeling in PostgreSQL for instantaneous margin-at-risk aggregation.
- Asynchronous FastAPI server with webhook support for multichannel ingest.
- Tactical executive dashboard in Tailwind CSS and Vanilla JS with telemetry console.
Related Professional Packs
Executive Decision Pack: S&OP & Supply Chain Risk
Modelos cuantitativos ejecutivos en Excel y plantillas de mitigación de obsolescencia bajo norma IEC 62402.
Get Pack