A comprehensive technical guide to the LightNote AI proposal evaluation algorithm. Learn how we quantify proposal quality across 5 essential dimensions, predict win probabilities, and generate sentence-level diagnostics.
When an audit is executed via /api/audit-proposal, our algorithm calculates sub-scores (0-100) across five critical categories:
Evaluates sentence complexity, readability grade, and structural flow. Flags dense jargon, run-on sentences, and ambiguous language that confuses stakeholders.
Analyzes whether the text emphasizes client outcomes or mere feature lists. Measures the ratio of client-focused benefits vs. vendor-focused self-descriptions.
Detects fuzzy, untestable commitments. Looks for quantified milestones, exact deliverables, SLA response times, and verified case study metrics.
Scans for grammatical precision, tone consistency, formatting harmony, and executive presence suitable for C-suite decision-makers.
Verifies the presence and adequate depth of all 8 core proposal sections: Executive Summary, Problem, Solution, Scope, Timeline, Pricing, Why Us, Next Steps.
Based on empirical analysis of historical proposal outcomes, your overall audit score maps directly to closing probability:
| Overall Score | Verdict | Estimated Win Rate | Action Recommendation |
|---|---|---|---|
| 0 - 59 | High Risk | < 25% | Major rewrite required; weak value proposition |
| 60 - 74 | Average | 35% - 50% | Needs stronger pricing clarity and specific metrics |
| 75 - 87 | Competitive | 55% - 70% | Polished; resolve 2-3 flagged sentences to excel |
| 88 - 100 | Winning | > 75% | Exceptional; ready for client delivery with high confidence |
In addition to aggregate grades, the audit engine returns a detailed list of line-level issues: