Coverage and Extraction Fidelity

Overview

The Coverage tab of Rule Intelligence answers two questions: what is tested and where are the gaps? and how faithfully did we extract requirements from your documents? It's a read-only dashboard — nothing here changes your project. Explore the capability at coverage gap detection.

Before you begin

• Open the workspace (project editor → Requirements tab, or the standalone Rule Intelligence hub → select a project), then click the Coverage tab.
• The panel is headed Coverage & Extraction Fidelity with the subtitle "Tests per endpoint × dimension × strategy, plus extraction quality per uploaded artifact."
• The dashboard loads automatically for the selected project; there are no controls to configure.

Step 1 — Read the summary badges

At the top of the coverage section are four totals:

Step 2 — Read the per-endpoint coverage matrix

Under Per-endpoint coverage matrix, each row is an endpoint (method + path). Columns are:

• Total — total tests on that endpoint.
• One column per dimension (the test dimensions present in your project, shown with underscores as spaces).
• One column per strategy that has at least one test.

Cells are heat-shaded green — darker means more tests relative to the busiest cell — so thin coverage is visible at a glance. Any endpoint with zero tests is shaded red and tagged no tests, making gaps obvious. If the project has no endpoints, the matrix shows "No endpoints found for this project."

Step 3 — Read the extraction fidelity panel

The lower Extraction fidelity section reports how completely requirements were captured from your documents, with three totals:

• Requirements — total extracted.
• Average fidelity — mean fidelity across requirements (or n/a).
• Low fidelity — count of low-fidelity requirements.

Two side-by-side tables follow:

• By artifact (worst first) — each uploaded document with its Reqs count, Avg fidelity %, and Low count (highlighted when non-zero). Worst documents sort to the top so problem sources are easy to find.
• Top missing structured fields — the structured fields most often absent from extractions, with a Missing in count. When extraction is clean it reads "Nothing missing — structured extraction is clean."

Fidelity matters because downstream test generation reads the structured constraint fields — a requirement with low fidelity may have been captured as prose but is missing the machine-readable parts the generator needs.

Step 4 — Act on the gaps

• For endpoints tagged no tests, open the endpoint and use Generate tests, or target specific gaps with Generate tests for gaps. See Generate endpoint test cases.
• For low-fidelity artifacts, tighten the source document (clearer, more explicit constraints) and re-run AI Parse from Importing and reviewing requirements, then re-check this tab.

Notes

• The dashboard reflects the current state — generate tests or re-parse documents, then revisit to see the matrix and fidelity update.
• Coverage counts tests that exist; it does not judge whether each test is correct. Pair it with the traceability matrix to confirm that requirements are actually verified.

Related articles

• Understanding Rule Intelligence
• Generate endpoint test cases
• Requirements traceability matrix
• Test Generation Settings