Skip to main content
Back to Help

Run a drawing review workflow

A drawing review applies your standards and design rules to a set of drawings and returns an issue list — typically a table of findings plus marked-up pages you can re-import into Bluebeam or your PDF tool of choice.

What you'll need​

  • The drawing set as a PDF (single file or a folder of PDFs).
  • Standards / specs / criteria the reviewer should apply (project specs, ASHRAE / NFPA, internal checklists, code references). Add these as context files on the workflow.
  • A scope — which disciplines, which pages, which severities — set in the workflow's instructions.

A good first review is small and focused: one discipline, one or two sheets, a clear checklist. Once that looks right, expand.

Start with the right standard​

The standard is what tells Nomic what "good" looks like. A template or example workflow can show the shape of a review, but it may not match your project type, jurisdiction, discipline, or stage of design.

For the most useful first run:

  • Use your own project standards, client criteria, internal checklist, or approved example comments when you have them.
  • Remove example standards that do not apply to the current drawing set.
  • Tell the workflow what not to flag, especially for early design packages where placeholders or unresolved details are expected.
  • If you do not have a formal standard yet, start from a small set of known review comments or RFIs and turn those into a checklist before reviewing a full package.

Run it​

  1. Open the drawing review workflow.
  2. Upload the drawing set into the Drawing set slot.
  3. (Optional) attach any one-off references for this specific review.
  4. Click Run.

Runs stream issues in as they're found — you don't have to wait for the whole set to finish to start reviewing.

What you get back​

A typical drawing review returns:

  • An issue table — page, location, discipline, severity, description, citation.
  • Markups on the source pages — leader lines pointing at the issue with a short comment.
  • A summary — counts by severity, by discipline, and any pages that were skipped.

Review the issue table and markups before sharing them. Accept findings that are useful, reject false positives, and keep unresolved items for a human reviewer when more context is needed.

Make the output more useful​

  • Tighten the instructions. Examples beat adjectives — "flag any door that swings into an egress path narrower than 44 inches" is more reliable than "flag egress issues."
  • Trim the standards. A few highly relevant standards work better than dumping every spec.
  • Pick a sensible severity model. "Critical, major, minor, informational" is plenty — reviewers tune out long severity scales.
  • Run a small sample first. A 5-page sample test takes 1-2 minutes and shows you whether the workflow understands the task before you spend on a full set.