Patent Opportunity Matching System

Overview

A biotech firm needed to know which patents in its portfolio could win Department of Defense SBIR/STTR and CSO solicitations, and to keep the underlying patent database clean enough to trust the answer.

The challenge

Matching patents to defense opportunities was slow, manual, and undefendable.

1. Evaluating a single match meant reading full patent claims against every solicitation's problem statement by hand.
2. Manual data entry bred duplicates and broken links between patents, inventors, and assignees.
3. There was no defensible ranking of which patent to submit against which opportunity, only opinion.

The approach

1. Stage one, retrieval and structuring: patents are pulled by the system, 10 per run, claims structured with numbers and dependencies, duplicates detected, and every patent linked to its inventors and assignees.
2. Stage two, opportunity matching: each candidate patent is scored against the solicitation on 6 weighted criteria: technical and capability match (30%), use context and application (20%), defense platform alignment (15%), operational suitability (15%), technology readiness level (10%), and dual-use potential (10%).
3. The weighted composite lands 0 to 100 and assigns one of 5 fit tiers, with per-criterion sub-scores showing exactly why a patent ranked where it did.

The results

1. 10 patents retrieved, structured, and linked per run, with zero hand entry
2. Every patent-solicitation pair scored 0 to 100 across 6 weighted criteria
3. 5-tier fit ranking with per-criterion sub-scores explaining every placement
4. Submission choices backed by scores instead of opinion
5. Entity linking and duplicate detection keep the patent database trustworthy
6. Data entry reduced from hours to minutes