How to Prioritize Drawing Coordination Issues by Severity

Prioritizing drawing coordination issues by severity means sorting every conflict in a set by what it puts at risk, life safety first, then cost and schedule, then trade sequence, so the design team fixes the dangerous and expensive items before the cosmetic ones. Helonic, an AI construction drawing analysis platform, rates every finding on a 2D PDF set high, medium, or low and pins it to a page location. This article covers the triage method a coordination lead applies on top of that first rating.

It sits next to two related pieces. Our preconstruction risk findings show where conflicts concentrate across 1,000+ reviews, and RFI quality vs. RFI volume covers how to write the question once you know which issue matters. This one is about the order you work the list.

A raw issue count tells you nothing about what to fix first

A full-set review on a large job returns hundreds or thousands of items. On an 800-sheet data center campus, Helonic listed more than 2,100 issues for an owner's rep team of two people with two weeks. The 156 floor-loading conflicts in that list were estimated at $890,000 in change orders if they had surfaced after the slab pour. Read in sheet order, those 156 would have been scattered among sheet-index typos.

BIM coordination tools learned this years ago. Navisworks sorts clash results by severity by default, and Solibri's own clash detection guide gives the example that a small pipe through a bearing wall ranks below a large duct crossing a major beam. A PDF review needs the same discipline, and it needs a method that holds up when two people disagree.

Question 1: Does it touch life safety or code?

If yes, it starts as high and stays high until the design team resolves it. These are the issues that can stop a permit, fail an inspection, or hold up occupancy regardless of what they cost to fix.

• Egress. Corridor and stair widths, door swing, travel distance, and exit counts under IBC Chapter 10.
• Fire-rated assemblies with unprotected penetrations, or a rated wall on the life-safety plan that the architectural plan draws as non-rated.
• Sprinkler coverage blocked by something another discipline drew. The data center set had 89 heads sitting below hot-aisle containment panels.
• Accessibility clearances at doors, toilets, and counters.
• Anything in the structural load path, such as a beam landing on a wall another sheet shows as open.

Helonic's fire and life safety checks and code compliance category cover most of this list on the PDFs, and every hit carries its sheet and location so a reviewer can confirm it in a minute.

Question 2: What does it cost if it is found in the field?

Price the field fix, not the drawing fix. Almost every coordination issue is a cheap edit on paper. What separates high from medium is what has to be torn out, re-ordered, or re-poured if nobody catches it until the work is in place.

On a custom home in the Miami area, a finished floor elevation that disagreed with the flood-zone site plan was estimated at about $25,000 in pour and grading rework if found after slab work started. A beam landing on a wall drawn as open great room was about $12,000 after framing. Both were one-line fixes before construction. A useful rule of thumb: structure and anything embedded in it outranks overhead MEP, which outranks finishes.

Question 3: Which trade builds first?

Sequence can promote an issue even when the dollar figure is small. A sleeve missing from a slab pour scheduled for next week has to be answered this week. A ceiling conflict on a floor that will not see drywall for four months can wait for the next revision.

1. Below grade, foundations, and underslab utilities
2. Structure, slab penetrations, and embeds
3. Risers and overhead MEP mains
4. Branch ductwork, piping, and conduit
5. Ceilings, partitions, and finishes

Overlay that order on the project schedule and a medium item on step 2 often belongs above a high item on step 5. The coordination lead makes that call. Software does not know your pour dates.

Question 4: Who owns the fix?

An issue with a clear owner moves. An issue between two consultants sits until someone decides whose sheet changes. Tag every high and medium item with the design firm that owns the sheet, and pull the two-owner items into the next coordination meeting as decisions, not RFIs.

On a 340-sheet mixed-use tower drawn by four firms, the critical items went into Procore as RFIs tagged to the firm that owned each sheet, and the team issued 127 RFIs in one batch instead of over several weeks. The clash detection report guide covers the owner and status fields that make this tracking work.

A three-level severity matrix you can use on Monday

Three levels are enough. Five-level scales tend to collapse into arguments about the difference between a 3 and a 4.

Apply questions 1 and 2 to set the level, then use questions 3 and 4 to order items inside each level. That keeps life safety at the top without letting a low-cost slab sleeve fall behind a ceiling clash.

How do Helonic's severity ratings feed the triage?

Helonic does the first sort so people spend their time on questions 3 and 4. It reads every sheet of the 2D PDF set and gives each finding a severity (high, medium, or low), one of ten categories from coordination and code compliance to dimensions and QA/QC, and exact page location coordinates. The 2D clash detection findings and the code compliance findings are the place to start working the high bucket, since structure-MEP and life-safety conflicts live there.

• On a three-school K-12 bond package of more than 1,200 sheets, QA/QC sorted the 841 findings by severity and building, so senior staff reviewed only the locations that needed a correction.
• On the tower, the team confirmed 94% of 847 findings as real problems and dismissed the remaining 6% as low-confidence items.
• High items go to Procore or Autodesk as RFIs in one click through RFI generation, with the sheet reference attached.

Treat the rating as the first cut. A reviewer still adjusts it with what the drawings cannot show: the pour schedule, the contract, and decisions the owner already made. How AI handles borderline flags is covered in AI drawing review false positives, and the full review method is in how to review construction drawings.

Four triage mistakes that bury the real issues

• Working the list in sheet order. A-series problems get fixed first because they are on top, and the MEP conflicts wait.
• Letting the loudest trade set priority in the meeting.
• Counting issues per discipline and calling the biggest number the worst discipline. Fifty low items are less urgent than one high one.
• Never re-rating. A medium item becomes high when its trade starts shop drawings, and stale ratings stop meaning anything to the field.

If you are triaging a large set on a deadline, the 500-sheet PDF coordination plan puts this method on day two. For background on how AI review produces the list in the first place, see AI for construction drawings.