Design Audit: Evidence, Evaluation, and Limitations
Design Audit evaluates captured public web pages with AI-assisted rubrics. This reference explains the inputs, scope, and limits of that evidence—not the steps for using the app.
By Design Audit Team · Updated
What evidence does an evaluation use?
Inputs include captured page elements, page title and metadata, a screenshot when available, and available performance metrics. The selected evaluation uses configured criteria to analyze that evidence; the workflow consolidates its findings into a device-specific report.
The page workflow currently uses desktop captures at 1440 × 900 and mobile captures at 375 × 812. A capture represents one page state. A cookie overlay or incomplete load can affect the evidence, and these two sizes do not cover every responsive layout.
What can the findings establish?
Findings can identify conditions worth investigating in content, visual design, usability, accessibility, search, and performance. A reported score summarizes a rubric-based assessment; it is not a measurement of customer satisfaction or revenue impact.
A capture cannot establish complete keyboard or screen-reader behavior, a successful multi-page journey, search-engine indexing, or field performance across visitors. Use interaction review, user research, Search Console, or performance measurements according to the claim you need to investigate.
Who publishes this material?
Design Audit is a product by Recursive. The Design Audit Team byline identifies this product’s own editorial material, not an independent review. The current library was prepared with AI assistance using application source and the references linked from each article.
Examples are hypothetical unless explicitly identified otherwise. Update dates record content revisions. Corrections can be sent to [email protected] with the article URL and the statement concerned.