Columns Guide

This page explains what each results column means, where its value comes from, and how score is computed.

How the pipeline works
  1. Fetch and extract website content (title, headings, and paragraphs).
  2. Send extracted content to the selected AI model for structured analysis.
  3. Compute a deterministic final score from AI sub-scores and confidence.
URL / Title

Extracted by the scraper. The app fetches each URL, parses HTML with BeautifulSoup, reads the <title> tag, and builds extracted text from meta description, headings, and paragraphs (up to 12,000 characters).

Score (0-100)

Calculated locally, not generated directly by AI. Formula: base = (category_fit + buyer_match) / 6 * 90, then multiplied by 0.7 + 0.3 * confidence, and clamped to 0-100.

Confidence (0.0-1.0)

Returned by the AI model. It represents how confident the model is in its own qualification output, based on available page evidence.

Summary

AI-generated short description of what the prospect company appears to do.

Notes

AI-generated supporting context and caveats to explain the model's qualification judgment.

Intent Signals

AI-generated list of positive buying/fit indicators observed on the prospect website.

Red Flags

AI-generated list of negative indicators or risks that may reduce fit.