Preliminary design review (PDR) and critical design review (CDR) are two formal gates within the broader engineering design review process, which also includes requirements reviews, informal peer reviews, supplier reviews, test-readiness reviews, and production-readiness reviews.
Both are decision gates that a program has to clear to earn the go-ahead for its next phase. The terms trace back to aerosapce and defense systems engineering, where they were first formalized. They are now standard design reviews across manufacturing and hardware development.
A Preliminary Design Review comes early in detailed design, once the overall architecture and major interfaces are set but while the drawings are still cheap to change. It confirms the chosen approach can realistically meet its requirements and surfaces the risks worth resolving now, each with an owner and a plan, rather than after everything downstream has been built on top of them.
A Critical Design Review comes later, once the design is detailed and nearly complete. It checks the finished design against its requirements and constraints and confirms it's ready to move into production — the last structured point to catch a problem before manufacturing turns it into an expensive change.
The practical difference between the two review types involves maturity and stakes. PDR asks whether the design is worth developing further, while CDR asks whether the finished design is ready to build.
Using CoLab for PDR and CDR
CoLab runs both reviews in one place, with the option for AI to take the first pass. Before a design reaches the cross-functional team, AutoReview, CoLab's AI peer checker, runs against your own standards, guidelines, and lessons learned and annotates the drawing or 3D model with the errors and non-conformances a first review would catch, so reviewers walk into a cleaner design and spend their time on the tradeoffs that need judgment.
From there, design reviews can happen in the browser with no CAD install. Feedback is attached to the exact geometry it concerns, and every comment is captured automatically as a tracked issue with an owner and a status. That record is what keeps a problem raised at PDR from quietly resurfacing as a surprise at CDR. This product data is also what CoLab's AI agents draw on to surface lessons from past programs when a similar issue comes up again.