RFI Cycle Simulator
RL Framework · The RFI Doctrine

Routing-first RFI economics

Describe the program, then choose how far routing tightens. The simulator builds a queue that reproduces your current median, then reruns the same program with fewer hand-offs, a cap on how long each one waits, and fewer clarification loops. Review capacity is the same in both runs.

Reviewer load today iHow loaded the people who answer RFIs are right now. It cannot be read from the median, so it is your judgment. Comfortable is about 70 percent of review capacity in use, busy 85, stretched 95.
Routing discipline iHow consistently RFIs are triaged, assigned, chased, clarified, and closed through the right owner path. Low is the program as it runs today. Medium and High change the three behaviors listed under Assumptions.
RFI impact profile iFrames the consequence of unresolved RFIs. It changes the advisory interpretation, not the simulation.
Check it against your RFI log

Paste rows with two dates each: opened, then answered. Dates like 2026-03-14 are safest. The rows are read on this device and are not stored or sent.

Median cycle-time reduction 3.3 days per RFI · 12.0 to 8.7 days
Routing Index 56 / 100
Weak routing control. High wait-load exposure for field-blocking items.
25% to 36% answered inside 7 days
85% to 73% reviewer load
~19,000 RFI-days of waiting removed

Cycle time, days iElapsed days from submission to logged answer. The median is the typical RFI. The 80th and 95th percentiles are the slow tail, where field-blocking items sit.

Where the days go, mean days per RFI

NowTightened

Open RFIs over the program

NowTightened

What a change in RFI volume does iMean cycle time from the queueing formula as monthly volume moves, with the same reviewers. A line ends where arrivals outrun review capacity and the queue stops clearing.

NowTightened

Where to act first

    Read the RFI note
    Assumptions behind the simulation

    One RFI moves through four stages: routing hand-offs, a review queue with a fixed number of review slots, clarification loops, and closeout. Arrivals are random at the monthly rate. Stage times are random and right-skewed (exponential). The clock is calendar days.

    Before tightening, the mean cycle is split 36 percent routing, 24 percent clarification, 30 percent review, 10 percent closeout. That split and the levels below are the author's field estimates from mission-critical programs. They are not survey data, and no published study was found that measures these rates. One process study describes up to four handovers before a response (Pitman, Wijayaratna, and da Rocha, 2024), which is where Low sits. Medium and High are set so the saving on the default program matches the first version of this simulator. Tune them to your own log.

    What each routing discipline level assumes
    LevelHand-offs before the right reviewer has itLongest wait with one ownerResponses that come back for clarification
    LowAbout 4No cap40%
    MediumAbout 33 days33%
    HighAbout 22 days25%

    Planning aid. The simulation is calibrated to one number, the current median, and rests on the routing assumptions listed above. It is not a schedule forecast and makes no float-recovery claim.

    Foundations

    This simulator draws on queueing theory and on RFI and project-controls practice:

    • Queueing theory (Erlang C, the M/M/c queue): the review stage. The simulation engine is checked against this closed form.
    • Discrete-event simulation: 16 seeded runs per result, so the same link reproduces the same numbers
    • Navigant Construction Forum: The Impact & Control of RFIs on Construction Projects (1,362 projects, about 1.1 million RFIs, median reply 9.7 days)
    • CSI / AIA construction contract administration: the RFI process and response obligations (A201 General Conditions)
    • Lean Construction / Last Planner System: RFIs as lookahead constraints cleared before they block work

    Roy Lan · Assistant Professor of Construction Management · roylanuzoma.com