If you've managed active jobs at scale, you've seen the same pattern. Requests for Information (RFI) pile up fast. In the construction-phase context, that workflow is distinct from owner-issued Request for Proposals (RFPs) used during procurement.
Using the benchmark of 9.9 RFIs per $1M of contract value from a Navigant Construction Forum study of over one million RFIs across 1,362 projects, a $500M commercial project would imply roughly 4,950 RFIs over its lifecycle. At $1,080 in administrative cost per RFI, according to the same study, that implies roughly $5.3M in processing overhead before a single delay claim enters the picture.
The manual workflow of retrieving, cross-referencing, and synthesizing information across drawings, specs, and prior RFI rulings creates a serious scaling problem.
RFI vs RFP: Two Instruments, Two Directions, Two Phases
On active jobs, teams confuse these terms when they should not. In the construction-phase context, an RFI and an RFP serve different parties, often flow in opposite directions, and belong to different project phases.
What an RFI Does in Construction
A construction-phase RFI flows upstream, from the contractor to the design team. Its purpose is to resolve ambiguities or conflicts in the contract documents. That includes drawings that contradict specifications, incomplete details at field-installation scale, or interdisciplinary coordination gaps between structural, architectural, and MEP systems where ASHRAE standards and architectural details intersect.
The AIA G716 form is the standard instrument, and under AIA A201-2017 § 3.2.8, it is used on A201 projects unless the contract documents specify another form.
An RFI response clarifies design intent. It does not modify the contract, authorize additional cost, or extend the schedule. Any cost or time impact requires a separate Change Order.
What an RFP Does in Construction
An RFP flows downstream, from the owner to shortlisted contractors. A common sequence starts with a Request for Qualifications (RFQ) to establish qualifications, followed by an RFP requesting cost, approach, and schedule from qualified firms.
The Comparison That Matters
Dimension | RFI (Construction Phase) | RFP |
Direction | Upstream: contractor to designer | Downstream: owner to contractors |
Phase | Preconstruction and construction | Preconstruction only |
Standard Form | AIA G716-2004 | No single AIA form |
Purpose | Clarify document ambiguities | Solicit cost, schedule, approach |
Contract Effect | Does not modify contract | Leads to contract negotiation |
The Standard RFI Process from Field Question to Formal Closeout
This workflow is straightforward on paper and messy in the field. An RFI touches every major party on a project, and the contractual stakes at each step are real.
Subcontractors originate most RFIs because questions arise at the point of installation. The general contractor reviews, filters, and routes upstream. The architect is the primary responder under AIA A201 and routes technically specialized questions to the relevant engineer of record.
How the Workflow Moves
Before drafting, the contractor should confirm the answer is not already in the contract documents. Per AIA A201 §3.2.7, this creates exposure to architect cost deductions when the information is already available.
This is where Datagrid's Deep Search Agent executes search across specs, drawings, RFIs, and submittals to get accurate answers grounded in project requirements.
If the answer is genuinely absent, the contractor drafts the RFI with specific drawing and specification references, a clear description of the ambiguity, and a proposed resolution. The GC routes to the architect, who prepares the response, with consultant input where needed, and distributes back through the GC.
Once accepted, the RFI is formally closed and becomes part of the project closeout package.
CSI MasterFormat includes RFI-related sections in Division 01 and Division 00 that support project-specific procedures. ConsensusDocs 200 includes a 14-calendar-day written response provision for contractor claims.
RFI Tracking: Logs, Cycle Time, and Best Practices
If you don't track RFIs with discipline, response lag becomes invisible until it turns into a schedule problem. A standard RFI log tracks the same core fields:
Submission dates
Response dates
Responsible parties
Referenced project files
Status
Priority flags tied to schedule activities
The gap between "date response required" and "date actually answered" is a key cycle-time metric. According to the previously mentioned Navigant Construction Forum study, the median across the industry is 9.7 days.
Five Practices to Enforce on Every Project
One question per RFI. Mixed-issue RFIs cause confusion and extend response times.
Include one specific recommendation, not multiple options. When the design team responds "either option is acceptable," documentation stalls.
Use automated reminders and visible accountability tracking. Every party should know who currently holds the RFI.
Link RFIs to schedule activities and flag critical path impacts. Unresolved RFIs on critical path activities are a primary driver of delay disputes.
Define response timeframes by party type in contract documents. Set maximum windows by recipient before the project starts.
Building these practices into project setup turns the RFI log from a passive record into a leading indicator that surfaces risk while there is still time to act on it.
Where Manual RFI Tracking and Response Breaks Down
I see the same failure pattern on large jobs. Volume climbs, response windows slip, and project files get fragmented across systems. Teams spend more time finding information than interpreting it.
The Volume Problem
RFI volume scales with project size in ways that break manual capacity. The Navigant Construction Forum study analyzed over one million RFIs and found that volume scales roughly linearly with contract value, so larger projects generate proportionally larger backlogs.
An AIA contract administration guide cites estimates that a single RFI requires approximately 8 to 17 person-hours for architects and engineers to review and respond to properly. Multiply that across hundreds or thousands of RFIs per project, and you have a capacity crisis that no amount of overtime solves.
The Response Gap
The median response time carries real schedule implications, and 21.9% of RFIs in the Navigant study received no response at all. The same study identified the no-response rate as a primary dispute predictor.
CMAA's Delivering Dispute-Free Projects guide explicitly identifies slow responses to information requests as a primary mechanism for delay claims, citing scenarios where contractors need a 5-day response but owners take 30 days or more.
The File Fragmentation Problem
Answering an RFI is rarely a quick lookup. It requires real digging across the project record before anyone can draft a defensible response.
The respondent must locate the relevant specification section and compare it against the drawing detail. They must check whether a prior RFI already addressed a related question and determine which source takes precedence under the project file hierarchy.
That information sits in separate platforms. Drawings are in one place. Specifications are in email. Schedules are in another tool. Prior RFI responses are in a third.
On projects with distributed teams across time zones, the lag compounds. An FMI study found that 4 out of 5 contractors cite low-quality design documents as a top external factor stunting productivity.
What It Costs
The Navigant study calculated an average total processing cost of roughly $859,680 per project for RFI review and response alone, and that figure does not include the downstream cost of schedule slippage when responses arrive late.
An unanswered RFI eventually shows up as documented evidence in a delay or defect claim.
Where AI Agents Fit Into the RFI Workflow
The manual bottleneck is retrieval, comparison, and judgment across scattered project files. A human expert manually searches across multiple platforms, reads and compares content across drawings, specs, submittals, and prior RFIs, then synthesizes a defensible answer. McKinsey's technology outlook characterizes the agentic AI shift as "using AI not just for retrieval but also for reasoning enabled by faster knowledge generation that can be scaled."
That shift maps directly to RFI work. Instead of a project engineer spending hours pulling files from multiple systems before drafting a response, AI agents handle the retrieval and cross-referencing so people focus on the technical call.
FMI frames this as a competitive necessity in its AI thought leadership: "a pragmatic understanding of AI's capabilities and constraints, coupled with an openness to embrace innovation and think inventively, will be crucial for contractors to remain competitive and address some of the challenges facing the industry."
How Datagrid's Agents Apply to RFIs
Datagrid's RFI Checker Agent and Deep Search Agent are built around the parts of the workflow that consume the most time:
Resolve questions against the project record by cross-referencing specs, drawings, submittals, and prior RFIs before anything routes to the design team
Surface conflicts across documents so respondents draft answers against current, governing information
Keep the workflow moving with visibility into response status and outstanding items
When an RFI sits on top of conflicting project files, the Deep Search Agent flags discrepancies before drafting begins. A fast answer is only useful if it is grounded in the governing project files.
What Practitioners Are Seeing
The practical value is simple. Less time spent hunting through project files. More time spent on judgment.
Alain Waha, CTO at Buro Happold, describes the impact directly:
"Buro is deploying Datagrid to help our designers and engineers resolve RFIs. We are on our way to reduce CA work by 50%."
That outcome maps to the core problem Datagrid's AI agents are built to solve. By taking on the manual retrieval and cross-referencing work around each RFI, agents free design professionals to spend more time on technical interpretation and less on document archaeology.
If your team is working through an RFI backlog and scattered project files are eating into your team's schedule, see how Datagrid's RFI Checker Agent and Deep Search Agent cross-reference questions against your existing project record before they ever reach the design team.
Book a demo to walk through your specific RFI workflow with the Datagrid team.
