An AQL sampling plan tells you how many units to inspect from a lot and how many defectives you can find before rejecting it. The plan depends on the lot size, the inspection level and the acceptable quality limit (AQL), and follows industry-standard acceptance sampling practice for single sampling under normal inspection.
β³ Values are pre-filled with an example β replace them with your own data.
Your data
parts
Normal is the usual choice. Reduced takes smaller samples, heightened larger ones.
The acceptable quality limit agreed for this defect class.
Result
Sample size125
Units to inspect, drawn at random from the whole lot.
Accept the lot with up to3
Accept if the sample contains 3 or fewer defectives.
Reject the lot with4
Reject if the sample contains 4 or more defectives.
Sample-size code letterK
Code letter K for a lot of 3,000 units at this inspection level.
Accepting the lot with 3 or fewer defective units in a sample of 125 does not prove the lot is good. It means a process running at about the AQL will be accepted most of the time; an individual lot can still contain more defectives.
Decision for a sample of 125: 0 to 3 defectives accepts the lot (green), 4 or more rejects it (red).
Formula
1. Lot size + inspection level β sample-size code letter
2. Code letter β sample size n
3. Code letter + AQL β acceptance number Ac and rejection number Re (Re = Ac + 1)
4. If the plan points to a neighbouring code letter, use that letter's sample size and numbers
Decision: defectives in the sample β€ Ac β accept the lot; β₯ Re β reject it
Worked example
A distributor receives a lot of 3,000 plastic connectors and agreed an AQL of 1.0 % with the supplier at the normal level. How many should be inspected?
Data
β’ Lot size = 3,000 units
β’ Inspection level = normal
β’ AQL = 1.0 %
Step by step
13,000 falls in the 1,201β3,200 band β code letter K
2Code K β sample size 125 units, drawn at random from the whole lot
3At AQL 1.0 % code K gives Ac = 3, Re = 4
4Inspect all 125: 0β3 defectives β accept the lot; 4 or more β reject it
What it means
Inspect 125 connectors. Up to 3 defectives accepts the lot; 4 or more rejects it. Accepting at 3 does not mean the lot is 1 % defective or better β it means a process running at about 1 % defective will be accepted most of the time. A single lot can still contain more defectives than that, especially if they are clustered.
Tips & common mistakes
β²Draw the sample at random from the whole lot, not from the top of one box.
β²Agree the AQL, inspection level and what counts as a defective in writing before the first delivery.
β²Use different AQLs for different defect classes (for example critical, major, minor).
β²Record every lot result: switching between normal, tightened and reduced inspection depends on the history.
β²AI assistants often invent accept numbers or mix up code letters; check every plan against this calculator before using it.
AI prompt for this calculation
Paste it into your AI assistant to be walked through the calculation with your own data.
Act as a supplier quality engineer and help me set up a single-sampling inspection plan (normal inspection).
Work in this order:
1. Lot size + inspection level (reduced / normal / heightened) β sample-size code letter.
2. Code letter β sample size n.
3. Code letter + AQL β acceptance number Ac and rejection number Re (Re = Ac + 1). If the plan for that cell points to another code letter, use that letter's sample size and numbers.
4. Decision: defectives found β€ Ac β accept; β₯ Re β reject.
Rules:
1. Restate my inputs first and check the thousands separators.
2. If the lot size, inspection level, AQL or defect definition is missing or unclear, ask me β never assume.
3. Tell me how to draw a random sample from my lot.
4. Explain what accepting the lot does and does not guarantee about its quality.
5. Tell me what to do if the lot is rejected (containment, sorting or rework, supplier notification).
My data:
Lot size =
Inspection level =
AQL (%) =
What counts as a defective =
The buttons copy the prompt first, then open the assistant in a new tab.
Add this calculator to your site
Copy the code below into your page. The calculator stays free; keep the attribution link with it.
<iframe src="https://www.qualityforeveryone.com/embed/calculators/en/aql-sampling-calculator" title="AQL Sampling β Quality for Everyone" width="100%" height="1100" style="border:0;max-width:100%" loading="lazy"></iframe>
<p><a href="https://www.qualityforeveryone.com/en/lab/calculators/aql-sampling-calculator">AQL Sampling calculator by Quality for Everyone</a></p>
Next step Β· DOCAMIN
Generate the inspection instruction document for this plan in DOCAMIN.
Less than many people think. AQL is the worst process average that the sampling plan treats as acceptable over a series of lots. It does not guarantee that any single lot contains at most that percentage of defectives, and a lot with more defectives can still pass if they are not picked. It is a supplier agreement tool, not a quality promise for each lot.
What are reduced and tightened inspection?
+
Switching rules adjust the effort to the supplier's history. After a run of rejected lots you move to tightened inspection (stricter acceptance numbers). After a long run of accepted lots and a stable process you may move to reduced inspection (smaller samples). This calculator shows normal inspection, the usual starting point.
Single, double or multiple sampling β which should I use?
+
Single sampling is the simplest: one sample, one decision. Double and multiple sampling take a smaller first sample and only draw more when the result is unclear, which lowers the average amount of inspection but makes planning and training harder. For most incoming inspection, single sampling is enough.
What should I do with a rejected lot?
+
Contain it first: quarantine the lot and check whether earlier lots or stock are affected. Then decide with the supplier whether to sort 100 %, rework, return or scrap, and ask for a root-cause analysis and corrective action. Keep the records β they drive the switch to tightened inspection.
Last tested: October 2026 Β· Tested with: automated math tests (3+ cases) Β· Reviewed by: Quality for Everyone editorial team
