Construction quality control and ITP field guide

The QC program, and the ITP at the center of it

A construction quality control program is the set of planned, documented checks that builds quality into the work while it is being built, rather than inspecting it in after the fact. You define what good looks like, you check it at the right moments, and you prove it with records. The document that carries the whole thing is the Inspection and Test Plan, the ITP, which lists for each activity what gets inspected or tested, when, to what acceptance criteria, by whom, and where the hold and witness points sit that stop the work until it is signed off.

The program is not the binder on the shelf. It is four things done on every activity: build the ITP, enforce the hold points, document the nonconformances, and keep the records that prove the work met spec. Skip any one of the four and the program is decoration.

Quality starts before the first inspection and runs past the last one. The product gets approved at the submittal, which is covered in the RFI and submittal guide, and every inspection, test, and hold point released gets logged contemporaneously in the daily report, covered in the daily report guide. This guide is the middle of that chain: the ITP that plans the checks, the hold points that stop the work before a defect gets buried, and the NCR that handles what fails.

Quality is built in, not inspected in

You cannot inspect quality into a finished building. By the time the work is covered, poured, or energized, the quality is already in it or it is not, and an inspection at the end only tells you which. The defect gets found late, the rework is expensive, and the schedule eats it. Quality is built in by planning the checks before the work, holding them at the points where a defect would otherwise get buried, and recording the result.

This is the difference between a program that works and a punch walk that pretends to be one. The walkthrough at the end finds the scratched paint and the missing cover plate. It does not find the rebar that was short on cover, the weld that cracked under the fireproofing, or the equipment ground bonded to nothing, because that work is already hidden. The point of building quality in is to check those at the moment they are still visible and still cheap to fix.

How much of this any single project formalizes depends on the contract. The project specification and the QA/QC plan set the scope of the program, the codes set the mandatory inspections, and the owner and the AHJ set what they want to witness. The framing holds across all of them. Plan the quality in. Do not try to inspect it in.

What is the difference between QA and QC?

QA is the system. QC is the check. Quality assurance is the planning, the procedures, the roles, and the standards that make quality happen, written before the work starts. Quality control is the inspection, the testing, and the measuring of the actual work against those standards as it is built. QA is proactive and lives in documents. QC is hands-on and lives at the work.

A concrete example separates them cleanly. Writing the pour procedure, setting the acceptance criteria for slump and strength, and naming who signs off the rebar is QA. Measuring the slump at the point of placement, casting the cylinders, and inspecting the rebar before the pour is QC. The QA plan said what good looks like. The QC inspection confirmed the actual work met it.

You need both, and the people who run a program well stop arguing about the labels. The more QA you do up front, the less QC catches at the work, because the crew already knows the target and the procedure. A job heavy on QC and light on QA is a job inspecting defects it planned no way to prevent. The terms blur in everyday use, often written together as QA/QC, and that is fine. What matters is that someone planned the quality and someone independent checked it.

What is an inspection and test plan?

An inspection and test plan, the ITP, is the document that lists, for each activity, every inspection and test the work has to pass, in sequence, with the rule attached to each one. For each line it answers six questions: what gets inspected or tested, when in the sequence, to what acceptance criteria, by whom, what record proves it, and whether it is a hold point, a witness point, or a review. It is the quality roadmap for a feature of work.

Read across one row and you have the whole instruction. Activity: place structural concrete. Reference: the structural drawings, the spec section, and ACI 318. Method: measure slump, cast cylinders, inspect formwork and rebar. Acceptance criteria: the slump range and the 28-day strength on the spec. Frequency: per the sampling rate the spec sets. Responsible: the QC inspector and the special inspector. Record: the inspection report and the cylinder break reports. Point: hold point before the pour.

The ITP is built from the project documents, not from a template you carry between jobs. The activities come from the scope, the references and acceptance criteria come from the spec and the adopted codes, the hold and witness points come from the QA/QC plan and what the owner and AHJ choose to witness, and the special-inspection lines come from the code, commonly IBC Chapter 17 where it is adopted. A generic ITP that does not trace to the actual spec sections and drawings on your job is a form, not a plan. Confirm every line against the contract documents and the AHJ before you rely on it.

Building the ITP, column by column

Build the ITP by walking the activity from start to finish and asking what could go wrong and where you would catch it. Start with the activity list, broken down the way the work actually happens, feature by feature. For each activity, pull the reference: the spec section, the drawing detail, and the code that governs it. Those references are where the acceptance criteria come from, so get them right before anything else.

Then fill the rest of the row. The inspection or test method is how you check it, a measurement, a test, a visual against a detail. The acceptance criteria is the measurable pass-fail, taken straight from the reference. The frequency is how often or what sample, because checking the first one and none after is not control. The responsible party does the check and the verifier signs it, and on the lines the code or owner requires, that verifier is a third party. The record is the form that proves it happened.

The order of the rows is the order of the work. The ITP is read top to bottom as the activity proceeds, so the hold point before the pour sits above the pour, and the test before backfill sits above the backfill. Get the sequence right and the ITP enforces itself, because the next signature cannot be earned until the one above it is. Get it out of order and the hold points land after the work they were supposed to gate, which is no gate at all.

Acceptance criteria: define what good looks like

Acceptance criteria are the measurable standard the work has to meet, and the word that carries it is measurable. Not "looks good," not "to the satisfaction of the engineer" standing alone, but the tolerance, the test value, the dimension, the torque, the pass-fail number. A criterion you cannot measure is a criterion you cannot enforce, and it becomes an argument at the work instead of a decision.

The difference shows up the moment two people disagree. "Smooth finish" is a fight waiting to happen. "Floor flatness FF 35 measured per ASTM E1155" is a number a reading settles. "Properly compacted" is an opinion. "95 percent of maximum dry density per the Proctor, tested at the spec frequency" is a result. Write the criteria so the inspection ends in a number, not a negotiation.

The criteria come from the project specification and the adopted codes and standards, and where the spec is silent the engineer of record sets the standard. Do not invent a tolerance because you remember it from another job. Pull it from the spec section that governs the work, cite the standard the spec cites, and where the number is genuinely a judgment call, get the criterion agreed and written down before the work starts, not after it fails.

What is a hold point?

A hold point is a mandatory stop in the ITP where the work cannot proceed until the inspection is signed off. Not "should not." Cannot. The rebar before the pour, the welds before they get covered, the piping test before backfill, the bonding before the wall closes. Each is a hold point because once the next step happens, the work is buried and the defect is hidden, expensive, or both. The hold point catches it while it is still in the open.

This is the single most important control in the whole program. A hold point that is enforced means a defect cannot get covered without someone qualified looking at it first. A hold point that gets waived under schedule pressure is the exact mechanism by which bad work disappears into the building. The crew that is behind wants to pour, and the discipline of the program is the inspector and the contract holding the line until the inspection is done and signed.

Build the hold points into the schedule, not on top of it. If the rebar inspection is a hold point, the pour cannot be scheduled until the inspection slot is, with the notification time the contract requires ahead of it. Which inspections are hold points is set by the QA/QC plan, the spec, the code, and the owner, and the mandatory ones are not yours to waive. Confirm the hold points and the required notice on your project, and treat releasing one without the sign-off as what it is: building over an inspection that never happened.

Hold point vs witness point vs surveillance

The ITP marks each check by how much control it carries, and the three common levels are not interchangeable. A hold point is a mandatory stop: the work cannot proceed past it without the sign-off, full stop. A witness point is a notification: you tell the witnessing party the inspection is happening, they may attend, and if they do not show within the notice window the work may proceed. Surveillance, sometimes called review, is the inspector checking the work and the records at their discretion, without stopping anything.

The distinction that bites is hold versus witness. Miss a hold point and you have built over a mandatory stop, which is a real nonconformance no matter how good the work turned out. Miss a witness point after you gave proper notice and the witness did not attend, and you are usually clear to proceed, because the obligation was to notify, not to wait forever. Document the notice either way, because a witness point only protects you if you can prove you gave the notice the contract required.

Which check is a hold and which is a witness is set by the QA/QC plan, the spec, and the owner, and it varies by project and by activity. Do not assume. Read the ITP and the plan, confirm the notice periods, and where the contract is unclear, get it clarified before the work. The time to learn an inspection was a hold point is not after the concrete is in the forms.

The first-work inspection and the mock-up

Inspect the first one before you build a hundred. The first-work inspection, sometimes the initial inspection, is where you and the inspector and the owner agree what good looks like on the actual job, on the actual first piece, before the crew repeats it across the building. Get the first weld, the first rack lineup, the first slab section right and signed off, and it becomes the benchmark every following one is measured against.

The mock-up takes that further. It is a built-to-scale sample, a section of wall, a typical equipment skid, a representative rack and busway run, built and approved before the real work starts, so the standard for finish, fit, and method is a physical thing everyone can point to rather than an argument over the spec. On a data center where one data hall repeats across the building, the mock-up that catches a coordination or workmanship problem catches it once instead of in every hall.

Set the bar early, because the cost of getting it wrong compounds. A standard agreed on the first piece costs one inspection. The same standard discovered on the hundredth piece costs ninety-nine reworks or an argument about whether any of them have to change. Which activities get a first-work inspection or a mock-up is set by the spec and the QA/QC plan, so confirm the requirement, and where the spec calls for a mock-up, do not start the production work until it is approved.

Three-phase control: preparatory, initial, follow-up

The federal three-phase control model is the cleanest structure for running quality on a feature of work, and many private QA/QC plans borrow it whole. It comes from the US Army Corps of Engineers and runs three phases per feature: preparatory before the work starts, initial on the first work, and follow-up on the ongoing work. NAVFAC and other federal agencies use the same model, and it maps directly onto the ITP.

The preparatory phase is the meeting before the work: review the spec and submittals, confirm the materials are approved and on hand, check that the preceding work is complete and acceptable, and walk the crew through the acceptance criteria and the hold points. The initial phase is the first-work inspection, where you check the first representative piece against the criteria and set the benchmark. The follow-up phase is the recurring inspection of the continuing work to confirm it stays at the standard the initial phase set.

The model works because it front-loads the thinking. Most quality problems are designed in before the first piece is built, by skipping the preparatory review, starting on incomplete preceding work, or never agreeing what good looks like. The three phases force those questions to the front. Whether your contract requires the three-phase model or not, the sequence of plan it, set the benchmark, then hold the benchmark is the discipline worth keeping.

What is a non-conformance report?

A non-conformance report, the NCR, is the formal document raised when work does not meet the spec. It records what the requirement was, how the work failed it, where and when, and then it tracks the work through to a disposition and a close-out. The NCR is not a punishment and it is not optional paperwork. It is how a defect gets controlled instead of quietly fixed, hidden, or argued about later.

Raise it when the work fails the acceptance criteria, not when you feel like it. A failed test, a dimension out of tolerance, a weld that did not pass, a material installed that was never approved at submittal. The NCR segregates the nonconforming work so it does not get covered or built upon while the disposition is decided, gives it a number so it can be tracked, and ties it to the requirement it missed so the fix can be checked against the same standard.

The NCR is a cousin of the RFI, but they answer different questions, and confusing them muddies both logs. An RFI asks the design team a question before the work is built, covered in the RFI and submittal guide. An NCR records work that was already built and failed the spec. Sometimes the disposition of an NCR needs an engineering answer, and that question routes like an RFI, but the NCR itself stays open and tracked until the nonconforming work is actually corrected and verified. An NCR that gets a verbal fix and never gets closed on paper is the most common way a quality record falls apart.

NCR disposition: rework, repair, use-as-is, reject

Every NCR ends in a disposition, the decision on what happens to the nonconforming work, and there are four standard ones. Rework brings the work back to the original spec, the cleanest outcome. Repair fixes it to a usable condition that does not fully meet the original spec, which usually needs the engineer's acceptance. Use-as-is accepts the nonconformance with no change because the deviation does not affect form, fit, or function, which always needs the engineer's or owner's concession. Reject means the work is scrapped or sent back.

The one that gets abused is use-as-is. It is the cheapest disposition for the contractor and the most tempting under schedule pressure, and it is exactly the one the field does not get to decide. Use-as-is and repair are engineering dispositions. The engineer of record, and on some items the owner and the AHJ, decide whether a deviation can stay, because they own the design intent and the liability that rides on it. The field documents the nonconformance and proposes nothing as accepted until the responsible engineer signs it.

Tie the disposition back to the same acceptance criteria the work missed. A rework is not closed because someone fixed something. It is closed because the corrected work was re-inspected against the original criterion and passed. A repair is not closed because the engineer accepted a method. It is closed because the repair was done to that accepted method and verified. The disposition decides the path. The verification against the criteria is what actually closes it.

Root cause and corrective action

Fixing the one defect closes the NCR. Fixing the reason it happened keeps the next one from being built. Those are two different jobs, and a program that only does the first keeps writing the same NCR over and over. The disposition handles the defect in front of you. The root cause and corrective action handle the system that produced it.

Ask why it happened, not just what is wrong. The wrong concrete showed up, again, because nobody was reviewing the batch ticket before the truck discharged. The welds keep failing because the procedure was never qualified for the position the crew is working in. A few rounds of "why" on a recurring NCR usually land on a process gap, not a careless worker, and the corrective action is a change to the process: a mandatory batch-ticket check before any pour, a re-qualified weld procedure, a hold point that was missing from the ITP. Fix the system and the defect stops recurring.

The corrective action is also the learning the program carries forward. An NCR that closes with a real root cause and a process change makes the next phase of work better. An NCR that closes with "reworked, see attached" and no cause analysis fixes one weld and teaches the job nothing. On repetitive work like a data center, where the same activity runs across many halls, the corrective action on the first NCR is worth more than the rework, because it prevents the same defect in every hall after it.

The records are the proof

If it is not documented, it did not happen. That is not a slogan on a QC job, it is the rule the dispute gets decided on. The inspection that was done but never signed, the test that passed but whose report nobody can find, the hold point released by a nod instead of a signature, all of those are, for the purpose of proving the work met spec, the same as never having happened. The record is the proof, and the proof is the deliverable.

The quality record is the set of inspection sign-offs, test reports, NCRs and their close-outs, and the as-built quality documentation that together prove the work met the spec at every controlled point. It assembles into the turnover package the owner gets at the end, which is the evidence the building was built to the documents. A job that kept its records clean as it went hands over a package in order. A job that did not spends closeout reconstructing what it can and explaining what it cannot.

The quality record and the daily report feed each other. The inspection witnessed, the hold point released, the test that passed or failed all get logged contemporaneously in the daily report, covered in the daily report guide, which is the field-level record of who was there and what was accepted on a given day. The ITP sign-offs and the test reports are the formal quality proof, and the daily report is the day-by-day log that anchors them in time. Keep both, and keep them tied together.

Testing and test reports

Testing is the part of QC that produces a number instead of an opinion, and the report is what makes the number proof. Concrete cylinders broken at a lab, soil compaction tested in the field, welds checked by nondestructive examination, electrical gear acceptance-tested, equipment pressure-tested, all of it produces a documented result measured against the acceptance criteria. The test without the report is hearsay. The report tied to the ITP line and the acceptance criteria is compliance you can show.

A lot of testing is done by a third party for a reason. Material tests, special inspections, and the independent lab break the conflict between the people who built the work and the result that says whether it passed. The concrete cylinders go to an accredited lab, the welds get inspected by a certified inspector, the soils get tested by the geotechnical firm. Their report is more credible precisely because they did not pour the concrete or run the weld.

What gets tested, how often, and to what standard is set by the spec and the codes, and the test methods are usually named standards, ASTM and the trade-specific ones, that the spec cites. Pull the frequency and the acceptance value from the spec, use the test method the spec calls out, and confirm the lab and the inspectors carry the accreditation the project and the AHJ require. A test run to the wrong method or sampled at the wrong rate is a number that does not prove what you need it to.

The QC manager is independent of production

The QC manager does not report to the people building the work. That independence is the point of the role. If the same person is responsible for getting the work done and for checking whether it met spec, the schedule pressure that pushes production is the same pressure pushing the inspection to pass, and the check stops being a check. The conflict is structural, not a question of anyone's character.

On a job that runs quality seriously, the QC manager has a separate report line, usually to the project executive or the company's quality organization, not to the superintendent whose schedule the inspections gate. And the QC manager has the authority to stop the work. A hold point means nothing if the person enforcing it can be overruled by the person it is holding up. The authority to say the pour does not happen until the rebar is signed off has to be real, and it has to be backed by the contract and the company.

How the role is structured varies with the project and the contract. Some owners require a named, qualified QC manager independent of the production chain, with the resume submitted and approved. The federal model requires it explicitly. Confirm what your contract and QA/QC plan require for the role, its qualifications, and its authority, and where the contract is thin, set the independence up anyway. A QC function that answers to production is one that passes work it should have failed.

Why the checker cannot be the doer

The principle runs below the manager too, all the way down to the individual inspection. The person who checks the work is not the person who did it. The foreman who hung the busway is not the right person to sign that it met spec, because he is the one who answers for the schedule if it did not. Separate the doing from the checking at every level the program can afford to.

This is also why the third-party special inspector exists, and why owners hire their own QA representatives over the contractor's QC. Each layer removes a little more of the pressure to pass. The inspector with no stake in the schedule signs what he actually sees. Give an inspector a quota to hit or a boss whose bonus rides on the milestone, and you have built the pressure to pass right back into the role you created to remove it.

Special inspection and IBC Chapter 17

Special inspection is the code-required, independent inspection of specific structural and life-safety work, and it sits alongside the contractor's QC rather than replacing it. Where the International Building Code is adopted, Chapter 17 is where it lives, listing the work that requires special inspection, commonly structural concrete, structural steel and welding, masonry, soils and foundations, and the anchorage and bracing of critical systems. The approved agency inspects it and reports to the building official and the engineer of record.

The distinction that matters on a job is that special inspection is the owner's and the code's inspection, not the contractor's. It is performed by an approved agency the owner engages, independent of the contractor, and its reports go to the AHJ. The contractor's QC still inspects the same work to the ITP. The two are separate lines on the ITP for the same activity, and both have to be satisfied. A passed special inspection does not relieve the contractor of its own QC, and a passed contractor inspection does not substitute for the special inspection the code requires.

What requires special inspection, and who can perform it, is set by the adopted code and the AHJ, and it varies by jurisdiction, edition, and the project's risk category. The statement of special inspections on the permit documents lists what applies to your job. Confirm the required special inspections against the adopted code, the permit, and the AHJ, and build them into the ITP as hold or witness points so the work is not covered before the code-required inspection happens.

Punch list vs QC: catch it during, not at the end

The punch list is the leftover, not the quality program. It is the list of minor, mostly cosmetic items found near the end, the touch-up paint, the missing trim, the adjustment, that get cleaned up before final acceptance. It has a place. The mistake is treating it as the quality control program, as if walking the building at the end and writing down what is wrong is how quality gets managed.

QC catches the defect during the work, at the hold point, while it is still visible and cheap to fix. The punch list catches what is left at the end, when the structural and hidden work is already covered and the only things still reachable are the finishes. A program that relies on the end walkthrough finds the scratched wall and misses the un-inspected weld behind it. In-process inspection at the right moments beats end inspection every time, and a job that ran its ITP well shows up to the punch walk with a short list, because the real defects were caught months earlier.

There is a direct relationship between the two. The more the ITP and the hold points catch during the work, the shorter the punch list at the end, and the fewer the surprises that turn into NCRs after the work is covered. A long punch list full of substantive items is a sign the in-process control was thin, not a sign the punch walk was thorough.

Quality starts at the submittal

The first quality check is not at the work, it is at the submittal, before the material is ever ordered. Approving the product against the spec before it is bought and installed is front-end quality control, and it is the cheapest place in the whole program to catch a mismatch. A fixture or a piece of gear that does not meet spec is a markup on a cut sheet while it is under submittal review. It is a much harder problem once it is installed and the NCR is open.

This is where the submittal process and the QC program connect directly. The approved submittal is the reference the field builds to and the inspector checks against, so the ITP line for an installed product points back to the approved submittal as part of its acceptance criteria. The submittal workflow, the register, the review stamps, and the long-lead procurement that drives it, is covered in the RFI and submittal guide. The quality point here is simple: an inspection of installed work is only as good as the submittal that approved the product going in. Catch the wrong product at the submittal, not at the hold point.

The QA/QC plan ties it together

The QA/QC plan is the project-specific document that defines how quality will be managed on this job, and the ITP is one part of it. The plan sets the quality policy, the organization and roles including the independent QC manager, the procedures, the inspection and testing program with the ITPs, the NCR and corrective-action process, and the records and turnover requirements. It is the document the owner reviews and approves at the start, and the one the program runs from.

Write it for the project, not from a template. A QA/QC plan that does not name the actual specs, the actual hold points, the actual special inspections, and the actual people is a binder that satisfies a submittal and controls nothing. The plan that works is the one the QC manager actually uses to run the inspections, the one the foremen know sets their hold points, and the one the records get filed against. The contract and the spec set what the plan has to contain, so build it to satisfy that and to actually run the job, which are usually the same thing done well.

Quality is everyone's job, not just the QC manager's

The QC manager cannot inspect quality into a job that does not care about it. There are never enough inspectors to watch every connection, so the quality that gets built depends on the foreman and the crew doing it right when no one is checking. The program sets the standard and verifies it. The people build to the standard or they do not, and the difference is culture.

Do it right the first time is cheaper than every alternative. The rework, the NCR, the schedule hit, the argument over disposition, all of it costs more than building it right once. The crews that internalize that the hold point is there to protect them, not to slow them down, are the ones that hit it ready, with the work done and the inspection earned. The crews that treat QC as the enemy are the ones whose work keeps getting buried before the inspection and dug back up by an NCR. The QC manager owns the system. Everyone owns the quality.

Keeping the ITP, the holds, and the NCRs in one place

The QC record falls apart the same way the daily report does, by living in too many places. The ITP is a spreadsheet on a laptop, the inspection sign-offs are on paper in the trailer, the NCRs are in an email thread, the test reports are in a folder somewhere, and the photos of the covered work are on a foreman's phone. Each piece is real, and together they cannot be assembled into the turnover package without weeks of hunting.

The fix is keeping the field side of quality in one system, attached to the job and the location. FieldOS is the offline-first field tool a lot of crews use for exactly this, holding the inspection records, the photos tied to the hold point and the location, the NCRs and their status, and the test results together so the quality record assembles itself instead of being reconstructed at closeout. On a data center under construction, with no signal three levels down in a concrete deck, the offline capture matters, because the hold-point photo and the inspection sign-off have to be made where the work is and sync later. The tool matters less than the discipline, but holding the discipline across thousands of inspections spread over a dozen places is close to impossible.

Common mistakes

• Trying to inspect quality in at the end with a punch walk instead of building it in with planned, in-process checks.
• Running with no hold points, so the rebar, the welds, and the tests get covered before anyone signs them off.
• Writing vague acceptance criteria like "looks good" instead of a measurable tolerance, test value, or pass-fail number.
• Leaving NCRs open with a verbal fix and no documented disposition, root cause, or close-out.
• Letting the field decide a use-as-is disposition that belongs to the engineer.
• Putting QC under the people it inspects, so schedule pressure passes work that should have failed.
• Keeping no records, so the work cannot be proven to have met spec when the turnover package or a dispute needs it.

What to document

Every controlled activity produces a record, and the record is only useful if it ties the requirement, the result, and the sign-off together in one place. The table is the working set for a QC record on a large data center or commercial job. Match it to the QA/QC plan and the spec, and add the activity and area breakdowns your project runs on.

Field checklist

Standards and references

The program is governed first by the contract, not by a single code. The project specification sets the acceptance criteria for each activity and lists the required inspections and tests, and the QA/QC plan sets how the program runs, the roles, the hold points, the NCR process, and the records. Those two documents control, and they are project specific, so read them at the start and build the ITP to satisfy them rather than from a template.

The codes set the mandatory inspections. Where the International Building Code is adopted, Chapter 17 lists the special inspections and tests and who reports them to the building official. The test methods are usually named standards the spec cites, ASTM for materials and many field tests, and the trade-specific ones, ACI for concrete, AWS for welding, and the rest, applied at the frequency the spec sets. The three-phase control model from the US Army Corps of Engineers is the standard structure for federal work and a common one for private QA/QC plans. The exact requirements vary by adopted code edition, jurisdiction, and risk category, so confirm them against the permit, the AHJ, and the spec.

Three rules hold across all of it. Quality is built in by the ITP, not inspected in at the end. Enforce the hold points so the work is not buried before the sign-off. And document it or it did not happen, which means closing every NCR with a real disposition and root cause. The contract, the codes, and the engineer govern the specifics. These three keep the program from becoming a binder nobody followed.

Terms and definitions

The QC program carries a vocabulary that shows up across the spec, the QA/QC plan, and the inspection software, and using the terms precisely is part of keeping a record that holds up.

Quality control (QC)

The inspection, testing, and measuring of the actual work against the acceptance criteria as it is built

Quality assurance (QA)

The planning, procedures, roles, and standards that make quality happen, set before the work starts

Inspection and test plan (ITP)

The document listing, per activity, what is inspected or tested, when, to what criteria, by whom, the record, and the hold or witness point

Acceptance criteria

The measurable standard the work must meet, a tolerance, test value, or pass-fail from the spec or code

Hold point

A mandatory stop where work cannot proceed until the inspection is signed off

Witness point

An inspection the witnessing party is notified of and may attend; work may proceed if they do not, after notice

First-work inspection / mock-up

The inspection of the first piece, or a built sample, that sets the agreed standard for the rest of the work

Non-conformance report (NCR)

The document recording work that failed the spec, tracked through disposition, root cause, and close-out

Disposition

The decision on nonconforming work: rework, repair, use-as-is, or reject, with engineering dispositions set by the engineer

Three-phase inspection

The preparatory, initial, and follow-up control model from the US Army Corps of Engineers

Special inspection

Code-required independent inspection of specific structural and life-safety work, commonly under IBC Chapter 17

Related tools

Calculators and readiness checks for this work

Compare your options

FAQ

People also ask

• What is an RFI in construction?
• What goes in a construction daily report?
• What is commissioning in construction?
• What is the difference between QA and QC in construction?
• What are zinc whiskers?