An automated prospecting workflow that gathers public business leads, audits their websites for technical and UX issues, and generates personalized outreach email drafts using LLM APIs.
Engineered to replace manual site inspection and repetitive copywriting with a structured, automated audit pipeline built on Python, FastAPI, and Next.js.
Traditional outbound marketing for digital agencies requires hours of manual work: researching prospects, inspecting websites for errors, and drafting individualized messages. This automation platform coordinates scraping, site evaluation, and language model generation into an integrated workflow.
Outreach emails that rely on generic templates produce poor response rates, but manually evaluating dozens of business websites each day is prohibitively slow for lean sales teams.
The pipeline coordinates multi-stage data gathering with strict verification rules.
I designed and implemented the backend extraction engine in Python and FastAPI, authored the LLM prompt templates with JSON schema constraints, and created the Next.js review interface.
To maintain credibility in generated outreach, the system uses strict ground-truth prompting: the language model only receives verified issues returned by the audit script. If no tangible technical flaws are found on a website, the draft pivots to a general introductory format rather than fabricating problems.
Python / FastAPI
Structured LLM APIs
Multi-Point Analyzer
Next.js
We are here to help you grow without hassle.
No call centres. No runaround. Just experienced people ready to help.