Why Most Construction RFIs Are Preventable (And How AI Catches Them at Plan Review)

SheetIntel Team

A construction startup called Zero RFI recently raised $13.8 million on a single premise: that a project can run with zero requests for information. They're not wrong about the goal. But most RFI-reduction tools focus on responding faster once an RFI lands — better tracking, faster routing, cleaner logs.

That's the wrong end of the problem.

The better question isn't "how do we process RFIs faster?" It's "why did this RFI happen at all?" Because in most cases, the answer traces back to something that was already in the drawings on bid day — a coordination conflict, a missing spec reference, a scope boundary nobody defined. Those issues were findable. They just weren't found.

This piece covers where field RFIs actually originate, how to catch them during pre-bid plan review, and how AI makes that process systematic instead of dependent on who happens to be reviewing that week.

The Real Cost of a Construction RFI

Before getting into prevention, the math is worth understanding. A single construction RFI costs between $1,000 and $3,000 to resolve when you factor in project manager time, architect response, documentation, and potential schedule impact. On a complex commercial project, 200 RFIs is not unusual — that's $200,000–$600,000 in overhead before a single legitimate change order.

The cost multiplier from pre-construction to field is severe. An issue caught during plan review takes a clarification email and maybe a spec update. The same issue caught after the framer has already built it costs a tear-out, a field change directive, and whoever's carrying the coordination risk absorbs it. Industry data consistently shows field RFIs cost 3–5x more than pre-construction findings to resolve.

MEP systems alone represent 30–40% of commercial construction costs and are the single largest source of coordination RFIs. That number is the ceiling on what's preventable if you review thoroughly before the bid.

Where RFIs Actually Come From: The 5 Root Causes

Most field RFIs trace back to one of five categories — all of which are visible in the drawings if you know where to look.

1. Trade Coordination Conflicts

Ductwork routed through a structural beam. Electrical conduit competing for ceiling space with HVAC supply. Plumbing stacks that don't align with floor penetrations on the structural drawings. These conflicts exist because MEP engineers work in separate models that are rarely fully coordinated before the design is issued for bid. A ceiling plenum with 14 inches of clearance and three trades trying to route through it will generate RFIs. Every time.

2. Missing or Contradictory Specifications

Plans reference a detail — "see spec section 09 21 16" — but that section doesn't exist in the project manual. Or a spec calls for a product that contradicts what's shown on the drawings. Or Division 1 scope descriptions don't match the Division-specific specs on alternates. These gaps are invisible during bidding and guaranteed to surface during submittal review or in the field.

3. Scope Boundary Gaps

"By others." Two of the most expensive words in a set of construction drawings. Who furnishes and installs the equipment curbs for the rooftop units? Who installs the door hardware — the hardware contractor or the door contractor? Who brings the utility connection to the building? Scope gaps at trade boundaries are fertile ground for disputes, back-charges, and RFIs once the job is underway.

4. Dimension and Detail Conflicts

The architectural floor plan shows a wall in one location. The structural plan shows a beam in the same location at a different dimension. A door schedule lists a 3'-0" opening. The structural rough opening is 2'-10". These are caught during a thorough review — but a 300-sheet plan set reviewed under bid deadline pressure means something gets missed.

5. Incomplete Design at Bid Time

Deferred submittals. Performance specifications with no design detail. "Design-build" scope buried in Division 1 that shows up as a defined product later. These are structural problems with the design package, and while you can't fully prevent them, you can identify them during plan review and price them as bid clarifications or exclusions before the contract is signed.

Pre-Bid Plan Review Checklist: Catching RFI Triggers Before They Cost You

The following checklist is structured around the five root causes above. Work through it systematically on every plan set — or run it through an AI review tool that does it automatically.

Trade Coordination

  • Check all MEP routing against structural framing plans — beam depths, joist spacing, slab penetrations
  • Verify ceiling plenum clearances — HVAC duct + structural depth + sprinkler + electrical = known conflict zone
  • Confirm plumbing stacks and floor drains align with structural slab penetrations
  • Review equipment access clearances (36" service clearance minimum for major mechanical equipment)
  • Check rooftop unit locations against structural framing for adequate support

Specification Review

  • Every specification reference on the drawings exists in the project manual
  • Division 1 scope descriptions agree with Division-specific specs
  • Alternates on drawings match bid form descriptions exactly
  • Material specs don't conflict with what's shown (e.g., spec says EPDM, drawing shows TPO)
  • Deferred submittals are flagged and priced appropriately

Scope Boundaries

  • All "by others" language is identified and accounted for in the bid
  • GC/sub scope splits on utility connections are defined (who brings it to 5 feet of building?)
  • Equipment curbs, housekeeping pads, and supports are scoped to a trade
  • Fire/life safety scope is complete and attributed
  • Temporary utilities scope is defined

Dimensions and Details

  • Section details agree with plan dimensions at critical interfaces
  • Door and window schedule sizes match structural rough openings
  • Finish schedule covers all spaces shown in the reflected ceiling plan
  • Structural engineer stamp present on all structural sheets
  • Column grid matches between architectural, structural, and MEP plans

Manual Review vs. AI Review: What Actually Gets Caught

A senior estimator working an 8-hour plan review on a 200-sheet set will catch most of the obvious conflicts — the ones in the areas they focus on. What they won't systematically catch is everything across every sheet, every discipline, every reference. Human attention is finite. Fatigue is real after hour four of a plan review.

Issue Type Manual Review (8 hrs) AI Plan Review (SheetIntel)
MEP/structural conflicts Catches obvious ones; misses complex ceiling stacks Flags geometry conflicts across all trades systematically
Missing spec references Catches some; many deferred to PM Cross-references all drawing notes against spec index
Scope boundary gaps Experience-dependent, not systematic Flags all "by others" and undefined scope language
Dimension conflicts Degrades after ~2–3 hours of review Consistent quality across full drawing set
Coverage on 200-sheet set Full day; focused areas only All sheets reviewed in minutes

What Happens When You Miss It at Bid

Here's a scenario that plays out on commercial projects constantly. You're bidding a 12,000 SF tenant improvement. You review the drawings, submit your number, win the job. Three weeks into framing, the mechanical sub calls. The HVAC supply duct that serves the open office can't route where it's drawn — there's a 24" structural beam in the way that nobody caught. The duct needs to drop 18 inches below ceiling in that zone.

Now the sequence: the GC writes an RFI. The RFI goes to the architect. The architect has a 7-day response window in the contract. The mechanical sub puts a crew hold on that area while they wait. Seven days later, the architect issues an architectural supplemental instruction. The duct drops, which means the ceiling drops, which means the lighting layout changes. Electrical now has a coordination issue. Another RFI.

What would have been a one-line bid clarification — "duct routing in Zone B conflicts with beam at column line D4; please clarify routing and ceiling height at bid" — is now a 3-week schedule impact and a change order fight over who carries the cost.

This is what the Caribou Coffee plan review found: three scope conflicts that would have been change orders. Caught at plan review, they were bid clarifications. Same information, 100x less expensive to resolve.

How AI Plan Review Fits Into Pre-Bid Workflow

SheetIntel isn't a replacement for your estimator's review. It's the first pass — the systematic sweep that catches what fatigued human attention misses on a deadline.

The workflow: upload your plan set (any size, any format), and SheetIntel reads every sheet across architectural, structural, MEP, and civil drawings. It flags coordination conflicts with specific sheet and detail references, surfaces "by others" scope gaps, cross-references specification notes against the project manual, and outputs a structured findings report. Your estimator then focuses their time on the items that need judgment — not on manually hunting for things the AI already found.

The output isn't a score or a rating. It's a list of specific issues with locations — the kind of thing you'd put in a bid clarification letter or an RFI-prevention checklist before the job starts.

Run your next plan set through SheetIntel

First review is free. Upload your plan set and get a trade-by-trade scope gap and coordination conflict report in minutes — before you bid.

Try SheetIntel Free →

The Bottom Line

Zero RFI raised money because the industry knows RFIs are expensive and mostly preventable. The disagreement is about where to attack the problem. Managing and routing RFIs faster is valuable — but it's downstream. The leverage is upstream, at plan review, before the contract is signed and before the crew is in the field.

Not every RFI is preventable. Owner-driven changes, unforeseen conditions, and design evolution will always generate some. But MEP coordination conflicts? Missing spec references? Scope boundary gaps? Those are findable in the drawings. That's the systematic work that AI plan review makes scalable — the first pass that runs on every job, not just the ones where your best estimator happens to have time.