Three-Story Single-Family Residence — HVAC / Structural Plenum Clash

• 11 findings on a signed 63-sheet set
• $10K+ field overrun, inside estimate
• 2+ weeks of schedule slip

Rose Remodeling & Construction, LLC came to Flikt.AI with a specific question rather than a general interest in software: would an automated coordination pass surface anything the design team and the field reviewers had not already caught? The set was 63 sheets across architectural, structural, MEP and civil, already signed, sealed and out for bid - the point in a project where a coordination problem is still cheap to fix and almost nobody is still looking for one.

The deciding factor was that Flikt.AI reads the 2D PDF set directly. There was no BIM model on this project, no federated model to clash, and no appetite for building one late in design. The review had to work on the documents that already existed, read every discipline against every other, and return findings that cite the sheets they came from so the team could check each one against the drawing rather than take it on trust.

Conflicts sat between disciplines, where every individual sheet was correct on its own terms. The mechanical engineer sized the duct correctly, the structural engineer sized the beams correctly, and the architect dimensioned the section correctly.

No MEP coordination sections were drawn, so duct clearance was never checked against beam depth.

The problem was a clearance issue rather than a hard hit. The duct could pass through the plenum - it just could not do that and hold the architect's ceiling heights and clear the deepest beams at the same time. That class of conflict survives sign-off.

63 sheets across four disciplines is more pairwise checking than a bid-period review realistically covers.

The RFI log only captures what somebody thinks to ask about. This clash was never raised as an RFI, so it was invisible until the duct met the beam.

Coordination review was distributed across the design team and the field. Each consultant checked their own discipline, the set was redlined at design submission, shop drawings were reviewed by trade, and anything that slipped through was expected to surface as an RFI during construction.

There was no step in that process where the full set was read against itself. No MEP coordination sections were drawn showing duct clearance relative to beam depth, so plenum clearance was never checked as its own question. Sign-off was organized by discipline, which means a conflict living between two disciplines had no owner until the field found it - and on this project the field did not find it either.

Flikt.AI returned 11 findings on a single residential set, with a combined flagged exposure of $64,552-$147,517. Finding C005 was flagged pre-construction at $8,831-$20,831. The actual field cost came in at $10,000+ plus two-plus weeks of schedule delay - squarely inside the flagged range, which is the useful part: the pre-construction estimate was accurate enough to have justified acting on it.

The resolution was permanent. A field-built interior soffit was framed to clear the displaced duct route, lowering finished ceiling height in the affected rooms and adding framing, drywall, finish and re-inspection to a schedule with no slack in it. None of that was in the architectural set.

The economics are straightforward. Flagged exposure across all 11 findings exceeded the cost of the review by more than an order of magnitude, and C005 on its own paid for the pass several times over. The saving is the difference between resolving a duct-versus-beam clearance conflict on paper during the bid period and resolving it in the field after the duct is already installed.