How to Write a Construction RFI That Gets Answered in 48 Hours (with AI)
Most construction RFIs are too vague to answer quickly. The 4-element structure (binary question, cost of waiting, proposed answer, annotated attachments).
A construction RFI that takes two weeks to come back was almost always written in a way that made a fast response impossible. The architect or engineer of record isn't slow because they're indifferent — they're slow because the question was unclear, the cost of being wrong on the answer was unknown, and the response required them to research the question before they could answer it.
The framework below changes that. An RFI written in the four elements gets a binary answer back, on average, in 48 hours. The work is in the writing — making the question answerable, naming the cost of waiting, and giving the design team exactly what they need to respond without doing the requester's homework for them.
A note on scope. This article is general guidance for construction project managers, field supervisors, and subcontractors. It is not legal advice or contract interpretation. The RFI process for your project is governed by your contract documents; the structure below is a writing framework that operates within whatever RFI process your contract requires.
Key takeaways
- Most RFIs are slow because they're written as open-ended research requests, not as binary questions a designer can answer in a sentence.
- The 4-element structure: binary question (yes/no or A/B), cost of waiting (specific impact), proposed answer (contractor's view), attachments (annotated photos and references).
- The proposed answer is the highest-leverage element. Designers respond to confirm-or-modify requests much faster than to draft-an-answer requests.
- Annotated photos are worth ten lines of prose. Arrows, dimensions, labels — the field condition becomes self-evident.
- The cost-of-waiting line gives the designer a priority signal they can act on without a phone call. Be honest — exaggerated cost claims lose credibility on subsequent RFIs.
Why most RFIs are slow
Four patterns dominate slow-RFI projects:
- Open-ended questions. "Please clarify the structural detail at gridline B/3" requires the design team to read the plans, identify the issue, and educate the contractor. That's not an RFI; it's a research request.
- Buried context. The relevant background — what's in the drawings, what the field condition is, what the contractor has already concluded — is in the RFI's third paragraph or in attached photos with no annotation. The design team can't find what they need quickly.
- No cost of waiting stated. The design team has no way to know whether this RFI is blocking work today, blocking work next week, or merely informational. They prioritize accordingly — which often means "later."
- No proposed answer. A well-written RFI proposes a specific answer for the design team to confirm, deny, or modify. Designers respond to proposals faster than they respond to open questions.
Each pattern is fixed by structure, not by length. A 60-word RFI written in the four elements outperforms a 600-word RFI that buries the question.
The 4-element RFI structure
The Construction RFI Generator produces RFIs in this structure. The field PM supplies the facts; the tool handles the framing.
Element 1 — The binary question
State the question as a yes/no or A/B choice the design team can answer with a single sentence.
Bad: "Please clarify the conflict between the structural drawings and the MEP drawings at the second-floor mechanical room."
Good: "Confirm whether the structural beam at gridline B/3 (S-201) takes priority over the duct run shown on M-302, or whether the duct routing should be revised. Photos and dimensions attached."
The binary frame forces the design team to make a decision rather than to write an essay. It also forces the requester to do the diagnostic work before sending the RFI — which is exactly where the time savings come from on the project's whole-cycle clock.
Element 2 — The cost of waiting
Name the impact, in calendar days and project activities, of a delayed response.
Response needed by [date] to maintain the framing schedule on the north wing. A response after [date] will require remobilization of the framing crew at an estimated impact of [N] additional crew-days.
The cost-of-waiting line does two things:
- It gives the design team a priority signal they can use without having to call to ask.
- It creates a project record of what depends on the response, which becomes the basis for any later schedule-impact analysis if the response is delayed past the stated date.
Be honest about the cost. An RFI that claims a fictitious cost of waiting to push priority loses credibility on the next RFI. Designers remember which contractors are accurate and which are not.
Element 3 — The proposed answer
For most RFIs, the requester has a view of what the answer should be. State it.
Contractor's proposed direction: revise duct routing per attached sketch, maintaining the structural beam as drawn. This proposal preserves the design intent of the structural system and adds an estimated [N] hours of MEP coordination work, currently absorbed in the contractor's GC fee.
The proposed answer:
- Gives the design team a specific thing to confirm or modify. A confirm-or-modify reply is much faster than a draft-an-answer reply.
- Surfaces the contractor's view of the cost implications, which is the appropriate place for that view to be on the record.
- Demonstrates the contractor has done the work. The RFI reads as a partnership, not as a homework dump.
A few RFIs genuinely have no proposed answer — true field conditions the designer needs to interpret de novo. Those are rare. When they do come up, name the absence: "No contractor-proposed direction; this is a true design question requiring engineer of record interpretation."
Element 4 — The attachments and references
Close with the minimum set of references the design team needs to respond, organized for fast reading.
- Drawing sheets referenced, with revision numbers.
- Specification sections referenced.
- Field photos with arrows or annotations pointing to the specific condition (not just "see attached" — annotated).
- Dimensional information the design team needs (existing field condition vs. drawing dimension).
- Prior related RFIs, if any, by RFI number.
Annotated photos are the single highest-ROI item in this list. A photo with an arrow pointing at the conflict and a dimension labeled in red is worth ten lines of prose describing the field condition. The Construction RFI Generator does not draw the arrows; the field team does. The tool reminds you to.
What the structure looks like end-to-end
A complete RFI in this structure fits on one page. It reads like a small, professional, factual artifact. It demonstrates that the contractor has diagnosed the issue, has a view, and respects the design team's time.
The same RFI, written without the structure, often runs to two or three pages, takes longer to write, takes longer to read, and gets answered slower. Length is not effort. Structure is effort.
What this approach is not
A few honest limits:
- It doesn't replace the contract's RFI process. Some contracts specify forms, distribution lists, and response-time provisions. Follow them. The four elements live inside whatever process the contract requires.
- It doesn't shift design responsibility. The proposed answer is what the contractor would do if given the call. It is not a substitute for the designer's professional judgment, and it does not relieve the contractor of the obligation to build per the contract documents once the answer is issued.
- It doesn't fix bad upstream documentation. RFIs originate from conditions in the drawings, the specs, or the field. If the project is generating RFIs at a high rate, the upstream documentation is likely the issue and the RFI process alone won't solve it.
What AI does well
The Construction RFI Generator handles the writing layer: turning a field-discovered issue into a structured RFI with a binary question, a cost-of-waiting statement, a proposed answer, and a clean reference list. The PM supplies the facts and the proposed direction; the tool handles the framing.
What AI does not do:
- Diagnose the issue. That's a field call.
- Quantify the cost of waiting accurately. That's the PM's call, informed by the schedule.
- Propose the right answer. The contractor's view is the contractor's view, formed from the field and the drawings.
- Substitute for the contract's RFI process. Distribution, signatures, and response-time clocks are governed by the contract.
A practical adoption pattern
For a project transitioning to this RFI structure:
- Week 1 — train the field team on the four elements. The proposed-answer element is usually the biggest behavior change.
- Week 2 — run a sample audit of the last 10 RFIs sent on the project. For each, ask: would the four-element version have gotten a faster response? In most cases the answer is obvious.
- Week 3 onward — every RFI in the four elements. Track response times. The design team's average response time should drop noticeably; if it doesn't, look at which element is still missing.
A project with disciplined RFI writing tends to surface fewer disputes downstream — because the issues that would have become disputes were resolved fast, on the record, with the design team's specific agreement. The RFI log becomes a clean project history rather than a list of unresolved questions.
How to start
Take the next RFI you would have sent and rewrite it in the four elements using the Construction RFI Generator. Send it. Time the response. Do the same for the next ten. Most projects see the median response time drop substantially in the first month.
Next steps
- Construction RFI Generator — four-element structure end-to-end.
- Construction Change Order Generator — for when the RFI response triggers a change. See also Change Orders That Don't Get Disputed.
- Construction Progress Report Generator — for the owner-facing summary of RFIs and their cost-of-waiting implications.
- Construction PM Claude Plugin install guide — to run all of this from inside Claude.
Frequently asked questions
What is a construction RFI?
A Request for Information (RFI) is a formal written question from the contractor to the design team (architect, engineer of record) that documents a question arising from the contract documents, a field condition, or a coordination issue. The RFI's purpose is to obtain a documented response that becomes part of the project record and that can be referenced in subsequent change orders or disputes.
How long should a construction RFI take to be answered?
Industry norms vary by project, but a well-written RFI on a routine field condition should typically be answerable within 48-72 hours. Contracts often specify a response window (5-10 business days is common). RFIs that take significantly longer usually trace to one of two causes: the question wasn't binary, or the priority (cost of waiting) wasn't communicated.
What makes a construction RFI fast?
Four elements: a binary question the designer can answer in a sentence; a clear cost-of-waiting statement that lets the designer prioritize; a proposed answer that the designer can confirm, modify, or deny; and annotated attachments (photos with arrows and dimensions, drawing-sheet references, prior related RFIs). Each element shifts work upstream to the requester, where the field knowledge is.
Should I propose an answer in my RFI?
Yes, for most RFIs. The contractor's view of the appropriate resolution, given the field conditions, is valuable information for the designer. A proposed answer turns the RFI into a confirm-or-modify request that the designer can resolve faster. True design questions with no contractor-proposed direction are appropriate but rare — name the absence explicitly when it applies.
How do I annotate photos for an RFI?
Arrows pointing to the specific condition, dimensions labeled in red or another contrasting color, references to the drawing sheet and detail being discussed, and any field-measured dimensions that differ from the drawn dimensions. A photo with an arrow and a label is worth ten lines of prose describing the field condition. Most RFI software supports this in the markup tool.
Does AI write good construction RFIs?
AI handles the writing layer: turning a field-discovered issue into a structured RFI with binary question, cost-of-waiting, proposed answer, and reference list. The PM supplies the facts and the proposed direction; the tool handles the framing. AI does not diagnose the issue, quantify the schedule impact, or substitute for the contract's RFI process.
What's the difference between an RFI and a change order?
An RFI asks a question; a change order documents a change. RFIs often precede change orders — the answer to the RFI triggers a change, which is then documented separately under the contract's change-order process. The RFI's purpose is to obtain the response; the change order's purpose is to amend the contract.
This article is general guidance for construction project managers. It is not legal advice or contract interpretation. The RFI process for your project is governed by your contract documents.
Save hours every week with the Contractor AI Cowork Vault
40 skills with lien-and-license guards for estimates, change orders, and customer comms.
Related Guides
Construction Change Orders That Don't Get Disputed: An AI Drafting Framework
Most disputed change orders fail on procedural issues, not substance. The 5-section drafting framework that gets owner approval the first time.
How to Install the Construction PM Claude Plugin (Cowork & Code)
Step-by-step installation guide for the Construction PM Claude plugin from The AI Career Lab — works in both Claude Cowork (chat) and Claude Code (terminal). What you get, how to install, and your first run.
How to Write a Construction Change Order with AI in 2026
A practical walkthrough for writing defensible construction change orders with AI — the right structure, common dispute traps, and the free tools that handle it.