What are the three types of inspection in a quality plan?

The three stages are incoming inspection (checking purchased material when it arrives, before it enters production), in-process inspection (checking the product at key steps during production so defects are caught before the next operation adds cost), and final inspection (checking the finished product against its full specification before it ships). Each catches a different defect at a different cost.

What is AQL in an inspection plan?

AQL, the acceptance quality limit, is the worst tolerable percent defective a sampling plan is built to routinely accept. Using the lot size, an inspection level, and a chosen AQL, a standard like ISO 2859-1 or ANSI/ASQ Z1.4 gives a sample size and an accept/reject number. A stricter AQL is used for critical characteristics and a looser one for cosmetic defects.

How do you decide sample size for inspection?

First decide whether to inspect 100 percent or sample. Use 100 percent inspection when a defect is critical to safety or function, or the process is not capable. Otherwise sample, and derive the sample size from an AQL sampling standard using the lot size and inspection level. The sample size and accept number are chosen so the statistical risk of accepting a bad lot is controlled.

What is the difference between an inspection plan and a control plan?

An inspection plan tells inspectors how to inspect, the characteristic, stage, method, sample, and accept/reject rule. A control plan is broader: it covers every way a characteristic is controlled, including prevention like error-proofing and process control like SPC, and inspection is only one of those methods. Often the control plan references the inspection plan for the inspection specifics.

How often should a quality inspection plan be updated?

An inspection plan should be reviewed whenever the product, process, supplier, or a customer requirement changes, and whenever a corrective action affects one of its characteristics. A named owner is responsible for that review. The most common failure is a plan that still checks characteristics the design dropped, or misses ones that a customer complaint should have added.

How to Write a Quality Inspection Plan

Defining what to inspect, when, how, and to what criteria across incoming, in-process, and final inspection, with AQL sampling and a row-by-row template.

A quality inspection plan is a document that defines exactly what to inspect, at which point in the process, how to measure it, how many units to check, and the criteria for accepting or rejecting the result, across incoming material, in-process steps, and final product. It is the instruction sheet that makes inspection repeatable instead of dependent on who is holding the caliper.

Inspection without a plan drifts. One inspector checks five parts, the next checks one; one accepts a light scratch, the next rejects it; a characteristic that mattered last month quietly stops getting checked. A quality inspection plan removes that variation by writing the inspection down: the characteristic, the stage, the method, the sample, and the accept/reject rule. This guide covers what an inspection plan is, the three inspection stages it spans, how to set sample size and acceptance criteria with AQL sampling, how to write one step by step, and how it fits alongside the control plan.

What is a quality inspection plan?

A quality inspection plan is a structured specification of the inspection activities for a product, each row naming a characteristic, where it is inspected, the method and equipment, the sample size and frequency, and the acceptance criteria. Its job is to make sure the right things get checked, the right way, at the right point, every time, regardless of which inspector is on shift.

It is narrower than a full control plan. A control plan covers every way a characteristic is controlled, including prevention methods like error-proofing and process controls like SPC. An inspection plan focuses specifically on the inspection activities: the checks where someone measures a characteristic and judges it against a limit. In many operations the inspection plan is the detailed inspection instructions that a control plan points to, the control plan says "inspect the bore, n=5, /hour, accept per plan," and the inspection plan spells out exactly how.

The value is consistency and evidence. A written inspection plan means two inspectors reach the same decision on the same part, and it produces a record that proves inspection happened to a defined standard, which is what an auditor, a customer, or a recall investigation will ask for. Without the plan, inspection is a set of personal habits; with it, inspection is a controlled process.

What are the three inspection stages?

A complete inspection plan spans three stages, each catching a different kind of problem at a different cost. Skipping any one of them pushes its defects downstream to a more expensive place.

Incoming inspection checks purchased materials and components when they arrive, before they enter production. This is the cheapest place to catch a supplier's defect, a bad lot rejected at the dock never gets built into finished goods. Incoming inspection is usually sampling-based and scaled to supplier performance: a proven supplier earns reduced inspection, a shaky one gets tighter checks.

In-process inspection checks the product at key steps during production, so a defect is caught before the next operation adds cost. In-process checks monitor process stability within a run and stop a drifting operation from producing a whole batch of scrap. This is where the first good piece of a run matters, a first article inspection at setup confirms the process is right before it makes a thousand parts wrong.

Final inspection checks the finished product against its full specification before it ships. This is the last chance to keep a defect from reaching the customer, and it is usually where the agreed acceptance quality limit with the customer is applied. Final inspection catches what the earlier stages missed, but relying on it alone is the expensive way to run, by the time a defect reaches final inspection, all the cost of making it is already spent.

Incoming, in-process, and final inspection each catch different defects. The later a defect is caught, the more it has already cost to make.

How do you set sample size and acceptance criteria?

You set sample size and acceptance criteria by deciding, for each characteristic, whether to inspect every unit or a sample, and if a sample, how large and how many defects you will tolerate. This is where the inspection plan gets rigorous instead of arbitrary.

The default question is 100% inspection versus sampling. You inspect 100% when a defect is critical to safety or function and the cost of an escape is high, or when the process is not capable enough to trust. You sample when 100% inspection is impractical or the risk allows it, which is most characteristics. Sampling trades a small, quantified risk of accepting a bad lot for a large saving in inspection effort.

When you sample, the acceptance quality limit (AQL) is how you set the bar. ISO 2859-1 defines AQL as the quality level that is the worst tolerable process average when a continuing series of lots is submitted for acceptance, in plain terms, the maximum percent defective you are willing to routinely accept (ISO 2859-1). From the lot size, an inspection level, and a chosen AQL, the standard's tables give you a sample size and an accept/reject number: check this many units, accept the lot if defects are at or below the accept number, reject it above. The US equivalent, ANSI/ASQ Z1.4, works the same way. A stricter (lower) AQL for critical characteristics, a looser one for cosmetic, the plan sets the tolerance honestly and in advance.

Each inspection plan row forces six answers, ending in the accept/reject rule, set from the AQL in advance so the decision is a calculation, not a judgment call.

What does AQL sampling actually standardize?

AQL sampling replaces guesswork about how many to check with a defined, statistically grounded plan. ISO 2859-1 specifies an acceptance sampling system for inspection by attributes, indexed by the acceptance quality limit, with single, double, and multiple sampling plans selected from the lot size and inspection level (ISO 2859-1). The point is that an AQL of, say, 1.0% means the plan is built to routinely accept lots running at or below one percent defective and reject worse ones, the sample size and accept number are chosen so the math holds. The value of writing the AQL into the inspection plan is that the tolerance is agreed and documented up front, usually with the customer, rather than argued about after a questionable lot ships. It also makes inspection auditable: anyone can check that the sample taken matches the plan the standard prescribes for that lot size and AQL.

How do you write a quality inspection plan?

You write a quality inspection plan by choosing what to inspect by risk, deciding where and how to check it, and fixing the accept/reject rule before any part is made. The sequence:

List the characteristics to inspect, by risk. Pull them from the drawing, the customer spec, and the risk analysis. Classify each as critical, major, or minor, the classification will drive how tightly it is inspected. Do not inspect everything; inspect what carries risk.
Assign each characteristic to a stage. Decide whether it is checked at incoming, in-process, or final, or more than one. Catch each defect at the earliest, cheapest stage where it can be found.
Choose the method and equipment. Specify the gauge, instrument, or test for each check, and confirm it is capable for the tolerance. A visual check needs a defined standard, a boundary sample or photo, so "acceptable finish" means the same thing to everyone.
Set sample size and frequency. Decide 100% or sampling. For sampling, use an AQL and the ISO 2859 or Z1.4 tables to get a sample size and accept number from the lot size and inspection level. Tighten for critical characteristics, loosen for cosmetic.
Write the acceptance criteria. State the target, tolerance, and the accept/reject rule for each characteristic as a number, tied to its source. The rule should leave no room for the inspector to interpret.
Define what happens on a reject. Name the reaction: contain the lot, tag and segregate, notify, and route to disposition. Link it to your nonconformance and corrective-action process so a reject is handled, not just noted.
Assign records and an owner. State what gets recorded and where, and name who owns the plan and revisits it after any change to the product, process, supplier, or requirement. Records turn inspection into evidence; an owner keeps the plan matching the floor.

How is an inspection plan different from a control plan?

An inspection plan tells inspectors how to inspect; a control plan tells the organization how a characteristic is controlled, of which inspection is one method. The control plan is broader and sits higher, it can call for prevention (error-proofing), process control (SPC), and inspection, and it names the reaction plan for each. The inspection plan is the detailed how-to for the inspection portion. In a small operation they may be the same document; in a mature one the control plan references the inspection plan for the specifics.

The most common mistake in both is the same: writing the plan and never keeping it current. An inspection plan that still checks a characteristic the design dropped, or misses one a customer complaint added, is worse than none, it spends effort looking in the wrong place while producing a record that says quality is under control. In batch production this shows up as inspection tied to a product revision the floor stopped running two changes ago.

Inspection plans also fail quietly when the records go nowhere. If inspection results land on a clipboard that gets filed unread, a rising reject trend is invisible until it becomes a customer complaint. The value of inspection is the fast signal when something is drifting, and that requires the data to come back quickly enough to act on. When inspection checks are captured at the station and flow into one live layer, a reject trend at incoming or a failing in-process check is visible in the moment, to the inspector and the quality team at once. Harmony digitizes the checks operators and inspectors already run and connects them to the systems around the line in one live layer with no rip-and-replace of existing gauges or software. When CLS moved its production and quality logging off paper its inspection data became a signal the team could act on during the shift rather than a record filed after the fact. Write the plan by risk, set honest acceptance criteria, and make sure the results come back fast enough to matter.