Quick Summary
- AI plan review works at any design stage—30%, 50%, 80%, or 100% CDs
- Earlier catches are cheaper to fix: 1x during SD, 10x during CD, 100x during construction
- Establish a quality baseline at each milestone—track improvement across design phases
- Re-checks on updated plans available at reduced rates—incentivizes iterative QA
Most teams wait until construction documents are "complete" before running any QA review. By then, fixing issues is expensive and disruptive. The most effective firms check at every major design milestone—30%, 50%, 80%, and final CDs. AI plan review makes this practical, because checking a progress set takes the same effort as checking final documents.
The Cost of Waiting
The construction industry's "1-10-100 rule" applies directly to plan review:
What to Check at Each Milestone
Design Stage Review Guide
Focus: major coordination issues, program compliance, code-level feasibility. Are the occupancy classifications right? Do egress paths make sense? Is the structural system compatible with the architectural layout?
Typical findings: 20-60 issues. Cheap to fix.
Focus: cross-discipline coordination, specification alignment, detail development. Do MEP layouts conflict with structural? Are material specifications consistent? Are accessibility requirements addressed?
Typical findings: 50-150 issues. Moderate cost to fix.
Focus: completeness, full code compliance, all cross-references, schedule coordination. This is the "last easy chance" to catch issues before the document goes to permitting or bidding.
Typical findings: 100-400 issues. Important to catch now.
Focus: final verification, addenda incorporation, comparison against previous checks. Confirm all previous issues were resolved and no new issues were introduced.
Typical findings: 30-80 residual issues if earlier checks were done.
Can AI Check Incomplete Drawings?
Yes. AI plan review doesn't require complete construction documents. It reviews whatever is provided and identifies issues within the scope of what's been designed:
- At 30%, it focuses on high-level coordination and feasibility
- At 50%, it adds cross-discipline checks as more detail is available
- At 80%, it runs full code compliance and completeness checks
- Missing information is flagged as "not yet shown" rather than "error"—the AI understands context
Real Example: Catching Issues at DD Stage
A hospitality project was reviewed at 50% DD. The AI found that the architectural floor plan showed a commercial kitchen, but no mechanical exhaust system was designed for it. Catching this at DD—before the MEP engineer had started CD-level design—saved weeks of coordination rework that would have happened if discovered during permitting.
Tracking Quality Across Milestones
When you check at multiple stages, you build a quality timeline. This data is valuable for project management and consultant accountability:
- Issue count trend: Are issues decreasing with each submission? If not, the design team may need support.
- Resolution rate: What percentage of 50% issues were resolved by 80%?
- New issue introduction: Are new coordination problems being created as the design evolves?
- Consultant comparison: Which consultants produce cleaner documents at each stage?
Re-Check Pricing
Iterative checking is priced to encourage it. Re-checks on updated plans for the same project are available at reduced rates—because the AI already understands the project context and can focus on what changed.
Start Checking Earlier
Don't wait for final CDs to discover coordination issues. Upload your progress set at any stage—AI identifies what needs attention now, so fixes happen while they're still cheap.
Check Your Progress Set
Upload your current design—whether it's 30% SD or 80% CD. AI identifies issues at the level of detail available and helps you fix them before the next milestone.
Get Started · From $50