Agentic SystemsSTATUS: Working Prototype 1 0

AI Sales Lead Qualifier

Automated intent scoring and routing agent graph

Built with Python, LangGraph, and FastAPI, this system serves as intelligent middleware between inbound inquiry forms and CRM platforms. It parses unstructured lead communication, extracts key budget and decision-maker entities, deterministically scores intent, and updates CRM records in real-time.

Inspect GitHub Source Code

01 // THE BUSINESS PROBLEM & SITUATION

Operational Context

Sales teams were spending hours manually reviewing unqualified inbound web inquiries, leading to delayed response times for high-value prospects.

CLIENT / ENVIRONMENT: Designed as a scalable solution for B2B sales teams dealing with high-volume inbound marketing traffic.

02 // SYSTEM ARCHITECTURE & SOLUTION

What Was Built

I architected an agentic workflow using LangGraph that intercepts incoming webhooks, analyzes intent against a strict qualification rubric, and interacts securely with the target CRM API.

leads-qualifier-ai.sysAgentic Systems

VERIFIED ARCHITECTUREAI Sales Lead Qualifier

SYSTEMS_ARCHITECTURE_EXPLODER // LAYERED_STACKFULL-STACK SYSTEM DECOUPLING

Frontend Layer

Next.js 16 App Router, React 19, Tailwind CSS, TypeScript

API & Middleware

FastAPI, Pydantic Schema Validation, Rate Limiters

Orchestration

LangGraph Multi-Agent Supervisor & State Machines

Model Providers

OpenAI GPT-4o, Anthropic Claude 3.5, Vision AI

Data Stores

PostgreSQL (Prisma), Qdrant Vector Store, Redis Caching

Observability

LangSmith traces, Pydantic audit logs, Error retries

03 // ENGINEERING CONTRIBUTION & FEATURES

Exact Contribution

Architected the multi-agent supervisor graph in LangGraph
Developed the FastAPI backend middleware with Pydantic validation
Integrated CRM API webhooks with retry queueing
Designed the administrative monitoring dashboard

Key Capabilities Built

✓ Context-aware Intent Scoring

✓ Automated CRM Entity Extraction

✓ Real-time Representative Alerting

✓ Strict Pydantic Output Validation

DATA_PROVENANCE_TIMELINE // AUDIT_TRACETRACE_ID: #TR-992041

14:22:01.002Source Ingestion

Received webhook from lead submission formOK

14:22:01.045Schema Sanitization

Pydantic validated 12 extracted fieldsOK

14:22:01.210LangGraph Reasoning

Score intent = 0.94 (Tier A Lead)OK

14:22:01.350Human Audit Check

Confidence threshold passed (>0.85)PASSED

14:22:01.480CRM Commit

Created contact ID #884920 in PostgreSQLCOMMITTED

04 // EXECUTION PIPELINE

Step-by-Step Workflow

STEP 01Prospect submits an inquiry form.

STEP 02FastAPI receives the webhook payload.

STEP 03LangGraph agent extracts buying signals and evaluates criteria.

STEP 04System validates JSON schema before external dispatch.

STEP 05Qualified prospects trigger immediate representative notification.

05 // ENGINEERING LESSONS & OUTCOMES

Technical Challenges Overcome

• Preventing model hallucination during criteria scoring.
• Handling upstream API rate limits under sudden traffic spikes.

Measurable Outcomes

→ Reduced initial qualification response time to under 30 seconds.

→ Eliminated manual review for 80% of low-intent inquiries.

→ Provided 100% deterministic JSON schemas for CRM ingestion.

Lessons & Engineering Rules

"AI outputs must be validated by deterministic schemas before external API dispatches."
"Isolating agent reasoning steps from action execution prevents partial state corruption."

NEXT RELATED CASE STUDY

WhatsApp Automation Hub

View Study

PROJECT METADATA

ARSLAN'S ROLEAI Engineer & Systems Architect

YEAR & STATUS2025 // Working Prototype

CATEGORYAgentic Systems

TECHNOLOGY STACK

PythonLangGraphFastAPINext.jsPostgreSQLPydanticDocker

GITHUB METRICS

Stars: 1

Forks: 0

Verified GitHub Update: 8/2/2026

Need a similar solution?

If your business faces similar data, automation, or software bottlenecks, let's discuss your requirements.

Start a Conversation