A pre-shipment inspection checks a finished, mostly packed lot at the factory before it ships. An inspector draws a random sample against an agreed sampling plan, compares it with the approved reference sample and specification, and reports the findings so the buyer can release, hold or reject the lot. For knives, the checklist must cover the tip, edge, handle, marking and blade protection. It does not replace laboratory tests for steel, hardness or food contact.
What a pre-shipment inspection covers
By the time an inspector arrives, the order is finished. The steel is bought, the grinding, polishing and assembly are done, the cartons are printed and the vessel is usually booked. That is the character of this check: it verifies a completed lot, and every remedy after a poor result costs money, time or both.
The scope should be written into the order, because an inspection company will otherwise apply its own default checklist. For kitchenware and knives, a useful scope usually covers:
- Quantity produced and packed, carton count and quantity per carton against the packing list.
- Workmanship and finish compared with the sealed reference sample, not with a written adjective.
- Dimensions and weight measured by a stated method against the drawing and its tolerances.
- Function checks for anything that moves, locks, closes or cuts, using a method agreed in advance.
- Logo, artwork revision, retail labels, barcodes and any required country-of-origin marking.
- Edge protection, inserts and the retail pack, plus carton construction, shipping marks and carton weights.
- Any carton drop or transit check, but only where it was agreed in advance as part of the scope.
What it does not cover is anything that cannot be seen or measured on site with the agreed tools. Steel composition, blade hardness, coating composition and food-contact migration are laboratory or qualification questions. Keep them on a separate track with your product documentation.
Is this the same as a government-mandated PSI?
No. The phrase has a formal meaning in trade law. The World Trade Organization's Agreement on Preshipment Inspection covers inspections that verify quality, quantity, price or customs classification of goods on behalf of an importing government, whether contracted or mandated by it. A buyer-commissioned final random inspection is a private quality check under your contract with the inspection company. If your destination market does require a government inspection, book it separately; one does not stand in for the other.
A knife-specific inspection checklist
A general kitchenware checklist misses the things that make knives different: a sharp edge that can be damaged by handling, a point that can pierce its own packaging, a handle joint that can loosen, and a blade steel that can spot if it is packed damp. The inspection should look at each zone of the knife against the approved sample.
Swipe or scroll the table to compare all columns.
| Zone | What to check | How it is usually checked |
|---|---|---|
| Tip | Point intact and straight, no bend, chip or burr | Visual under good light, compared with the reference sample |
| Cutting edge | Continuous bevel along the full length, no chips, rolled sections or unground areas | Visual check under good light and magnification; never by running a finger along the edge. A controlled cutting test by trained staff only if the method was agreed beforehand |
| Blade faces and spine | Finish, grind lines, scratches, discoloration, rust spots, logo position and legibility | Visual at an agreed viewing distance and lighting |
| Bolster and heel | Gaps, weld or solder marks, sharp corners at the finger guard | Visual check; touch checks only on non-cutting surfaces |
| Handle and tang | Scale fit, gaps, rivet flushness, cracks in wood or composite, glue residue, handle secure | Visual plus a gentle twist and pull check, as agreed |
| Dimensions and weight | Blade length, overall length, spine thickness, weight | Calibrated rule, caliper and scale against drawing tolerances |
| Marking and labels | Steel and care marking as approved, origin marking, barcode | Compared with approved artwork; barcodes scanned |
| Blade guard and pack | Guard covers the full edge and tip, stays on when the pack is handled, insert holds the knife | Handling check by the agreed method |
How should knife defects be classified?
Every finding is placed in a class, usually critical, major or minor, and each class has its own acceptance number in the sampling plan. The definitions are yours to agree with the factory before production. The examples below are illustrative only; your own list should come from the product risk, your market and the approved sample. A critical finding is a safety matter and is not something to settle with a price allowance.
Swipe or scroll the table to compare all columns.
| Example finding | Illustrative class | Reasoning to agree |
|---|---|---|
| Tip or edge exposed through the guard or retail pack | Critical | A handling and user injury risk |
| Loose handle or cracked handle scale | Critical or major, by agreement | Could fail in use; the severity depends on the construction |
| Chipped or rolled edge section | Major | The knife does not perform as sold |
| Rust spots on the blade | Major | Affects use and customer acceptance |
| Wrong barcode or missing origin marking | Major | Can stop receipt or sale |
| Light scratch visible only at close range | Minor | Cosmetic, if within the reference sample |
Random sampling and the lot decision
Acceptance sampling only works if the sample represents the lot. The lot needs to be finished, packed, marked and counted, with cartons numbered and staged so that the inspector can select from anywhere in the run rather than the front row. The sealed reference sample, the signed sample approval record and the approved artwork should be on the table from the start.
The sampling plan comes from a named standard. Most inspection bodies work to ISO 2859-1, and ISO's page for ISO 2859-1:2026 confirms that the third edition was published in January 2026 and "replaces the previous edition from 1999 along with its amendments." The sample sizes and acceptance numbers sit inside the paid standard, so they are not reproduced here, and no AQL value is a universal default. The NIST/SEMATECH Engineering Statistics Handbook describes the AQL as "a percent defective that is the base line requirement for the quality of the producer's product", which is a level you choose for each defect class. Our AQL inspection guide explains how levels and limits interact.
- The factory confirms the lot is complete, packed and matched to the packing list, and the slot is fixed against that status rather than the vessel date.
- The inspector selects cartons at random across the whole lot and records their numbers.
- Units are drawn from those cartons up to the sample size the named plan requires.
- Each unit is checked against the zone checklist and the reference sample, and every finding is photographed and located.
- Findings are classified using the agreed defect list.
- The count in each class is compared with that class's acceptance number.
- The report states whether the lot conforms; the buyer then decides to release, hold or reject it.
Book the inspection against a stated completion and packing condition, confirm the date only when the factory reports that condition met, and leave working days before the cut-off for a result that is not a pass. For a first order or a product with a known finish risk, a during-production inspection early in the run is usually the cheaper place to catch a problem, because the remaining pieces can still be corrected.
What happens when a lot fails
A failed result puts the lot on hold. Release authority sits with the buyer, but only within the limits of product safety and the legal requirements of the destination market. Within those limits, what matters is choosing a route deliberately, with the cost and schedule effect of each in view, rather than defaulting to whichever route the vessel date forces.
Swipe or scroll the table to compare all columns.
| Route | What it involves | What to watch |
|---|---|---|
| 100 percent sorting | Every piece checked, defectives removed, the lot repacked and the shortfall made up or shipped short | Works only for visible, separable defects; the time depends on lot size and the defect |
| Rework | Regrinding, repolishing, replacing a component or reprinting packaging | A wrong artwork revision on printed boxes means new print lead time |
| Concession | The buyer accepts an agreed deviation, usually against a price allowance, recorded in writing | Only for non-safety deviations that still meet every mandatory requirement; never for critical defects |
| Rejection | The lot is not shipped; remake, replacement or cancellation under the contract | Raises deposit, material and liability questions the contract should anticipate |
A sorted or reworked lot should be resubmitted as an identified lot, not presented as a fresh one, or the inspection history stops meaning anything. ISO notes that its scheme includes safeguards for consumers such as "switching to tightened inspection if issues are found", which is another reason to keep the history intact. Sorting also has limits: a dimension that has drifted across the whole run cannot be sorted out, because there is nothing good left to keep.
Who pays for a re-inspection?
There is no default; it depends on what the order says. Where a failure is clearly the factory's, the factory paying is a common position. Where it comes from an ambiguous specification or a late change, that is harder to argue. Settle it in the RFQ and carry it into the purchase order, because after a failure both sides negotiate under a running vessel cut-off.
Inspection versus laboratory testing
A visual and functional inspection cannot establish a steel grade, a hardness value, a migration result or a market compliance status. An inspector can confirm that a marking is present and matches the artwork. They cannot confirm that the claim it makes is true. Those questions need their own tests, scoped and completed before production where possible.
Swipe or scroll the table to compare all columns.
| Question | Pre-shipment inspection | Laboratory or qualification test |
|---|---|---|
| Is the blade the specified steel? | Checks the marking and the certificate reference | Composition analysis and certificate traced to the lot |
| Is the hardness on target? | Only with a calibrated tester and an agreed method | Hardness testing of sample blades to an agreed method and range |
| Does the edge cut and keep cutting? | Visual edge check; a simple cut check if agreed | A controlled sharpness and edge-retention test |
| Is it acceptable for food contact? | Cannot decide | Market-specific testing and documentation |
| Will the pack survive transit? | Checks pack-out, guard fit and carton marks | A transit test procedure matched to the route |
For cutting performance there is a published method. ISO's page for ISO 8442-5:2004 describes a test of the sharpness and edge retention of knives for food preparation, in which the blade makes controlled forward and reverse cutting strokes through a pack of synthetic test material. The related ISO 8442-1:1997 covers requirements for cutlery for the preparation of food. ISO lists both as confirmed in 2021 and expected to be replaced by new editions now at final draft stage, so ask the laboratory which edition it applies. Tests like these belong in sample qualification, not on the inspection table.
The report itself should be traceable: the lot and carton numbers, the standard and edition, the inspection level, the sample size and acceptance numbers used, each finding with a photograph and location, and a clear result. Keep it with the quotation, the sample approval and the packing list. For steel grade evidence, our guide to 304, 430 and blade steels sets out a verification sequence.
Your checklist
- Inspection scope written into the order, not just a date and an AQL value
- Knife checklist covering tip, edge, blade faces, bolster, handle, marking and blade guard
- Defect classes agreed in writing, with examples, before production
- Sampling standard and edition named, with an AQL chosen for each defect class
- Lot finished, packed, carton-numbered and staged for a random draw before the slot is fixed
- Sealed reference sample and signed approval record on site for the inspector
- Release authority and concession procedure agreed in the purchase order
- Re-inspection costs and the number included agreed in advance
- Sorted or reworked lots identified as such on resubmission
- Working days left before the cut-off for a result that is not a pass
- Steel, hardness, cutting and food-contact tests kept on a separate qualification track