AI Competitor Intelligence & Battlecard Generator
An automated competitor analysis system using SpyFu API, GPT, and Claude to generate structured PDF battlecards and dashboards for marketing teams.
Client: Digital Marketing Agency (Confidential)
01 Overview
Marketing and sales teams at growth-stage companies spend dozens of hours per month manually researching competitors - pulling ad data, analyzing messaging, and assembling battlecards for the sales team. This client needed the entire workflow automated: input a competitor domain, receive a structured competitive intelligence report within minutes.
We built an end-to-end pipeline that queries the SpyFu API for ad intelligence, uses LLMs to analyze messaging patterns, and generates polished PDF battlecards and interactive dashboards automatically.
02 The Challenge
• SpyFu API returns raw ad data - extracting strategic messaging signals and positioning patterns requires LLM-based interpretation, not simple parsing
• Battlecards need a consistent, opinionated format that sales reps can scan quickly - requiring careful prompt engineering and output structuring
• Some competitor analysis requires multi-source fusion: SpyFu ad data, LinkedIn job postings, and public pricing pages provide different intelligence signals
• PDF generation needed to handle variable-length content elegantly without layout breaks
• The client wanted a self-service dashboard, not just a file download
03 Our Solution
• **Multi-source intelligence pipeline**: SpyFu API provides ad history, keywords, and estimated ad spend. A scraping layer pulls public pricing and feature pages. LinkedIn API (where authorized) surfaces hiring signals. All sources feed a unified context builder.
• **LLM battlecard generation**: GPT analyzes messaging themes, unique selling propositions, and apparent target audiences. Claude handles objection handling suggestions and head-to-head positioning. Outputs are structured JSON before rendering to ensure consistent format.
• **ReportLab PDF generation**: Template-based PDF generation with dynamic sections, charts, and consistent branding - handles variable-length content without layout issues
• **Interactive dashboard**: FastAPI backend serves a React dashboard where users can request new battlecards, view historical reports, and export in multiple formats
• **LangChain orchestration**: The full multi-source retrieval + analysis pipeline orchestrated as a LangChain chain with retry logic and partial result caching
04 Results & Impact
Project Details
- Category
- AI/ML
- Client
- Digital Marketing Agency (Confidential)
Technologies Used
Related Projects
AI Accountability Coach - Multi-Channel Coaching System
A LangGraph multi-agent coaching system on GCP with Twilio and Make.com integration, delivering personalized reminders and motivational coaching across multiple channels.
Read case study →AuditMind: Multi-Agent Document Intelligence, From Days to Minutes
A scalable asynchronous multi-agent system that audits construction documents using parallel AI agents, real-time result streaming, and a production-grade backend.
Read case study →GraphRAG Knowledge Graph Consistency Checker
A multi-tenant Knowledge Graph system using GraphRAG + Neo4j AuraDB for Swedish construction tender audits, with full MLOps lifecycle from design to production.
Read case study →