The short answer
A single request for information (RFI) costs roughly $1,080 to review and respond to — counting only the labor to process the question and route the answer.¹
That number is almost beside the point. The processing cost is the tip of the iceberg. The RFI itself is usually a symptom of a documentation gap — a conflict, omission, or ambiguity in the drawings or specs — and the gap is what generates the expensive consequences downstream: rework, change orders, and schedule delay.
On a typical commercial project, the documentation-error bill runs into the low single-digit millions of dollars, and the majority of it is preventable before construction starts.
Here’s how that math works, with every number sourced.
How many RFIs is “normal”?
The most-cited industry benchmark comes from the Navigant Construction Forum’s 2013 study of more than 1,300 projects: about 9.9 RFIs per $1 million of construction value on a typical project.¹
That scales fast:
| Project value | Expected RFIs (≈9.9 / $M) | Processing cost (≈$1,080 / RFI) |
|---|---|---|
| $5M | ~50 | ~$54,000 |
| $25M | ~250 | ~$270,000 |
| $50M | ~500 | ~$535,000 |
| $100M | ~1,000 | ~$1,070,000 |
So before a single change order is written, a $50M commercial project can expect to spend roughly half a million dollars just processing RFIs — the back-and-forth of writing the question, researching the answer, and documenting the response.¹
Two things move that number:
- Documentation quality. The single biggest lever. A clean, well-coordinated drawing set generates far fewer RFIs than one full of conflicts and gaps.
- Project complexity. Healthcare, labs, and heavily-serviced buildings run hotter because they pack more disciplines into the same space.
The cost most people miss: open RFIs eat your schedule
The average RFI stays open for about 9.7 days waiting on a response.¹ Most of those don’t matter — but the ones sitting on the critical path stall the trades behind them. A handful of critical RFIs can quietly add weeks of cumulative slip to a job, and schedule delay is its own line-item cost (extended general conditions, escalation, liquidated damages).
This is the cost that never shows up cleanly on an invoice, which is exactly why it’s underestimated.
Change orders: the 5–10% you budgeted (and shouldn’t have to)
Most owners carry a contingency for change orders because they expect them. Industry baselines put change orders at roughly 5–10% of contract value on commercial work.²
On a $50M project, that’s $2.5M–$5M — money the owner has effectively pre-committed to fixing problems found during construction. A meaningful share of those changes trace back to the same root cause as the RFIs: information that conflicted, was missing, or was wrong in the documents.
Rework: where bad data becomes a hard cost
Rework — tearing out and redoing work that was already built — runs about 4–6% of total project value on average.³ And in a landmark FMI / Autodesk study, roughly 14% of all rework was attributed directly to bad data — wrong, missing, or conflicting information.⁴
For scale, a 2004 U.S. government study (NIST/NIBS) estimated that inadequate interoperability — the cost of information not lining up across the parties and systems building a facility — costs the U.S. capital-facilities industry about $15.8 billion a year.⁵
Putting it together: the documentation-error bill
Stack the layers for a representative $50M commercial new-build:
| Cost layer | Estimate | Source basis |
|---|---|---|
| RFI processing labor | ~$535K | 9.9 RFIs/$M × $1,080/RFI ¹ |
| Schedule slip (critical-path RFIs) | weeks of delay | 9.7-day avg open duration ¹ |
| Change orders | $2.5M–$5M | 5–10% of contract ² |
| Rework | $2M–$3M | 4–6% of project value ³ |
Not all of these are independent (a change order often pays for the rework), so you don’t simply add every row. But the order of magnitude is unambiguous: the documentation-error cost of a mid-size commercial project is measured in millions, not thousands — and a large portion of it traces back to a single, addressable cause.
The root cause: drawing and spec conflicts
When you sort RFIs and change orders by what triggered them, the largest single category is the same one every time: coordination conflicts and gaps across the drawing set — a duct routed through a beam, a fixture with no power shown, a dimension that disagrees between two sheets, a detail referenced but never drawn.
The reason these slip through is structural, not careless: every individual sheet can be correct while the set is wrong. The architect’s plan is right. The structural plan is right. The mechanical plan is right. The conflict only exists between them — and no single discipline owns the space where two trades overlap. Traditional plan review is done discipline-by-discipline, so the cross-discipline conflicts are exactly the ones that survive to the field.
That’s the gap that turns into an RFI, then a change order, then rework.
What you can actually do about it
The cheapest RFI is the one you never have to write, because the conflict was caught and resolved on paper. Three moves that move the needle:
- Treat preconstruction QA as a cross-discipline activity, not a per-discipline one. The conflicts live between the drawings, so the review has to look between them.
- Benchmark your RFI burn-rate. If you’re running well above ~10 RFIs per $1M, the documents — not the field — are usually the problem.
- Catch conflicts before the set is issued. Every conflict resolved in design is an RFI, a change order, and a rework cycle that never happens.
Where Flikt.AI fits
This is the problem Flikt.AI was built to solve. Flikt.AI is AI-powered plan review that cross-references complete 2D PDF plan sets across nine disciplines — architectural, structural, mechanical, electrical, plumbing, fire protection, civil, landscape, and interior design — to flag the coordination conflicts that become RFIs and change orders, before construction begins. No BIM model required.
On a South Florida commercial retail project, every conflict Flikt.AI flagged was confirmed real by the general contractor — zero false positives — representing six figures of avoided exposure. On a residential project, it caught an HVAC-vs-structure clash that the field team missed; resolving it after the fact cost $10,000+ and two weeks of schedule.
If the math above looks like your projects, that’s the cost Flikt.AI is designed to take off the table.
FAQ
How much does one RFI cost?
About $1,080 to review and respond to, on average, counting processing labor (Navigant Construction Forum, 2013).¹ The larger cost is usually the rework, change order, or schedule delay the RFI signals.
How many RFIs should a project expect?
Roughly 9.9 RFIs per $1 million of construction value on a typical project — so ~500 on a $50M job (Navigant, 2013).¹ Cleaner documentation lowers it; complex, heavily-serviced buildings raise it.
What percentage of a contract goes to change orders?
Industry baselines put change orders at about 5–10% of contract value on commercial work (CII / Dodge SmartMarket).²
How much does rework cost on a construction project?
About 4–6% of total project value on average, with roughly 14% of that rework caused by bad data — wrong, missing, or conflicting information (Dodge, 2018; FMI/Autodesk, 2020).³,⁴
What causes the most RFIs?
Coordination conflicts and gaps across the drawing set — information that disagrees or is missing between disciplines. Every sheet can be individually correct while the combined set still conflicts.
Can you prevent RFIs?
Not all of them, but a large share are preventable by catching drawing/spec conflicts before the set is issued — which is the purpose of cross-discipline preconstruction plan review.
Sources
- Navigant Construction Forum, Impact and Control of RFIs on Construction Projects, 2013 — RFI rate (9.9/$M), average cost per RFI ($1,080), average open duration (9.7 days).
- Construction Industry Institute, RT-265, Change Management, 2011; Dodge Data & Analytics, SmartMarket Report — change-order rate baseline (5–10% of contract).
- Dodge Data & Analytics, SmartMarket Report on Project Rework, 2018 — rework cost as % of project value (4–6%).
- FMI & Autodesk, Harnessing the Data Advantage in Construction, 2020 — share of rework attributable to bad data (~14%).
- NIST GCR 04-867 (NIBS), Cost Analysis of Inadequate Interoperability in the U.S. Capital Facilities Industry, 2004 — macro interoperability loss ($15.8B/yr).
Figures are drawn from publicly summarized study findings and are presented as benchmarks and order-of-magnitude ranges, not project-specific estimates. They are not suitable for litigation, damages, or contract negotiation.