Shopify work glove cut resistance schema for AI agents: ANSI/ISEA 105 A1-A9, EN 388 coupe vs TDM-100, and cross-standard confusion

1. The scale confusion: ANSI A4 ≠ EN 388 4 — different machines, incompatible numbers

Work glove cut resistance is tested and certified under two primary standards in the US and European markets. Both standards use numbers to report cut resistance. Both happen to use numbers in overlapping ranges. The numbers are not equivalent and cannot be compared, but they look like they should be.

ANSI/ISEA 105-2016 uses the TDM-100 blade-on-flat test machine. A straight blade is mounted to a load cell and drawn across a flat glove coupon — a small square of the actual glove material — under increasing load. The test reports the load in grams required to cut through 20 mm of material. Nine levels:

EN 388:2003 uses the Coup test machine. A small circular blade is rotated against the glove material under a fixed 5-Newton load while the sample moves in the opposite direction. The test measures how many cycles the rotating blade takes to cut through the sample versus the same number of cycles on a reference material (cotton canvas). The ratio — sample cycles divided by reference cycles — is reported on a 1–5 scale. A ratio of 2.5 or higher earns level 5; below 0.2 earns level 1.

The AI agent failure mode

A safety manager asks an AI shopping agent to find "cut level 4 or higher work gloves for sheet metal handling." The agent queries the catalog and filters products where the cut level field contains "4" or higher. It returns a result set mixing ANSI A4 gloves (1,500–2,199g TDM-100), EN 388:2003 level 4 gloves (coup ratio 10–19, unknown TDM-100 equivalent), and EN 388:2016 D gloves (3,000–3,999g TDM-100 — equivalent to ANSI A6). These products are presented as equally suitable for the application, ranked by price or review count.

The actual cut resistance of these three "level 4" products may vary from approximately ANSI A2 to ANSI A6 equivalent — a range spanning three ANSI levels and representing over a tenfold difference in cutting force (500g to 3,999g on the TDM-100). Sheet metal handling at high production volume requires A4 or higher on the TDM-100 scale; a glove that tests ANSI A2 equivalent but shows "EN 388 4" from the coup test is inadequate.

2. EN 388:2003 coup test vs EN 388:2016 TDM-100: same standard number, incompatible results

EN 388 is the European standard for protective gloves against mechanical risks — published by CEN (European Committee for Standardization). It covers cut, abrasion, tear, and puncture resistance. It is the standard most frequently cited on work gloves sold globally. It was revised in 2016. The revision is important enough to break all pre-2016 cut resistance ratings for one specific material class — HPPE/Dyneema — and yet many product listings still show "EN 388" without a year, or show numbers derived from the 2003 coup test as if they are current.

The coup test and its known failure mode for HPPE

The EN 388:2003 Coup test — also called the Coup test or TDM test in old documentation — uses a rotating circular blade pressed against the glove material. The blade spins at a fixed rate; the sample moves in the opposite direction; the test measures how many cycles it takes to cut through. The result is compared to cotton canvas reference material cycles. Higher ratio = more cut-resistant.

This test works adequately for conventional materials like leather, cotton, and standard synthetic fabrics. For HPPE (high-performance polyethylene, sold commercially as Dyneema, Spectra, and similar brands), the coup test produces systematically unreliable results due to frictional heat generation. The rotating blade makes repeated contact with the HPPE fibers, and HPPE's relatively low melting point (approximately 130–150°C for UHMWPE) means that the fibers at the blade contact point partially melt. In some cases, the melted material re-bonds and the blade "drags" rather than cuts cleanly, producing an inflated coup ratio — HPPE gloves testing at EN 388:2003 level 4 or 5 on the coup test while delivering only ANSI A2–A3 equivalent protection on the TDM-100. In other cases, the melted material causes the blade to pass through more easily, producing a deflated coup ratio — also wrong in the opposite direction.

Reading the EN 388 certification code

EN 388 certifications appear as a multi-digit code on the glove packaging and product listing. The code structure changed between the 2003 and 2016 editions:

A glove showing "EN 388 4X4X" was certified under the 2003 standard — the X's are masked here for position reference but represent actual numbers. The "4" in position 2 is the coup test result. No TDM-100 data exists for this glove unless it was retested. A glove showing "EN 388 4141F" was certified under the 2016 standard — the "F" suffix is the TDM-100 result (5,000+ grams, equivalent to ANSI A8). The coup test result is still in position 2 (here "1") but is supplemental.

3. Cut level alone is incomplete: puncture, abrasion, and impact are separate axes

ANSI/ISEA 105-2016 reports four distinct mechanical resistance properties, each tested independently and each reported on its own scale. A glove's ANSI cut level says absolutely nothing about its puncture resistance, abrasion resistance, or impact resistance. These properties are structurally independent: the materials that excel at cut resistance frequently perform poorly at the other three axes.

The four ANSI/ISEA 105 mechanical performance axes

Why high cut level frequently means low puncture and abrasion resistance

The materials that achieve ANSI A7–A9 cut ratings — HPPE/Dyneema fibers, sometimes combined with steel or glass fiber wrapping — are specifically engineered to resist lateral blade shearing forces. These same properties work against them in other test modes. HPPE fiber is smooth and slippery; a perpendicular probe slides between fibers rather than being resisted by them, producing low puncture resistance (P1 or P2) on gloves that test A7–A9 for cut. The same smooth fiber surface provides low friction when abraded against rough materials — the fibers slide and fray rather than resist, producing abrasion levels of 1–2 on TDM-100-A7 gloves.

Leather, by contrast, is structurally heterogeneous — the collagen fiber network in cowhide provides resistance to perpendicular probe penetration (P3–P4), sustained abrasion (level 4–6), and reasonable tear resistance. Its cut resistance on the TDM-100 is typically A1–A2: the collagen fibers offer minimal resistance to a sharp blade shearing laterally.

Glass handling: the puncture failure that cut level obscures

Handling flat glass sheets — window glass, automotive glass, display panel glass — exposes workers to two distinct injury modes: slicing along a clean edge (cut hazard) and penetration from a fragment point (puncture hazard). When a tempered glass pane shatters, it produces thousands of small cuboid fragments with blunt-ish but still penetrating points. An ANSI A7 HPPE glove prevents the edge slice but the fragment points punch through the low-puncture shell, lodging in the palm.

Professional glass handlers use gloves rated A6+ for cut and P3+ for puncture simultaneously — typically a hybrid construction with a cut-resistant knit shell (HPPE + glass fiber) and a leather palm overlay that provides P3–P4 puncture resistance. A product listing that shows only the cut level allows an AI agent to recommend a pure HPPE A8 glove for glass handling — which will pass the TDM-100 cut test and fail immediately on glass fragment puncture in real use.

4. Material tradeoffs: when an ANSI A2 leather glove outperforms an ANSI A7 HPPE glove

Material selection determines the full performance profile of a work glove. An AI agent that routes entirely on ANSI cut level will systematically select the wrong material for applications where cut is not the primary or sole hazard. Three applications where lower cut-level materials outperform higher cut-level alternatives:

Wire rope and cable handling

Wire rope handling exposes a worker's palm to sustained, repetitive friction as the rope moves through the grip. The rope's steel wires abrade the glove surface continuously. HPPE shell gloves rated ANSI A7–A9 have abrasion resistance levels of 1–2 (fails at 100–500 Martindale cycles). A rigger pulling 200 feet of wire rope through their hands in one operation may apply thousands of abrasion cycles against the palm in a single task. The HPPE shell wears through within hours — sometimes within a single heavy-use day — and the worn zone exposes bare hand skin to the rope surface.

Cowhide leather gloves rated ANSI A1–A2 for cut have abrasion resistance levels of 4–6 (2,000+ Martindale cycles). The leather palm outlasts the HPPE glove by a factor of 5–10× for wire rope work. After the HPPE glove has worn through in the cut-resistance zone, it no longer delivers A7 protection — the worn leather glove delivering A1 cut still delivers A1 cut consistently. For wire rope handling, durability of actual protection over a shift matters more than the peak cut level that exists only when the glove is new.

Welding-adjacent and heat exposure

HPPE and Dyneema fibers are polyethylene — thermoplastics with melting points of 130–150°C. Welding spatter (molten metal droplets) contacts the glove at 1,400°C or higher. Contact with welding spatter melts HPPE gloves through on first contact, burning through the shell and potentially fusing the melted fiber to the skin. Leather welding gloves (ANSI A1–A2) char rather than melt — the char layer provides momentary thermal insulation and the structural integrity of the leather is maintained up to approximately 200°C.

An AI agent asked for "cut-resistant gloves for a welding shop" that filters on highest ANSI cut level will recommend HPPE gloves. These are contraindicated in any environment where welding spatter, open flame, or sustained heat contact is present. The ANSI cut rating does not carry heat resistance information unless encoded separately in the product schema.

High-abrasion assembly and material handling

Workers handling rough castings, concrete masonry units, fiberglass panels, or abrasive granular materials apply abrasion force to the glove palm rather than cutting force. The correct performance axis is ANSI abrasion level (Martindale cycles to failure), not cut level. Reinforced leather or mechanics gloves (ANSI A2–A3 cut, abrasion level 5–6) outlast HPPE gloves (ANSI A7, abrasion level 1–2) by a factor of 4–10× in these environments. Durability of protection over a shift is the relevant metric.

5. Application routing failures: sheet metal, glass, box cutters, wire rope

The most consequential AI agent failures in work glove selection occur when a customer describes their application in plain language and the agent maps it to a single performance axis — cut level — when the application requires multiple axes or a different primary axis entirely. Four documented failure patterns:

HVAC sheet metal fabrication

HVAC ductwork fabrication involves cutting, bending, and assembling sheet metal — primarily galvanized steel in 26–18 gauge (0.5–1.2 mm thickness). The edges of cut sheet metal are razor-sharp. Workers handle panels continuously, sliding them through bending brakes, inserting them into fittings, and fastening. The cut hazard is real and significant: NIOSH occupational injury data shows sheet metal work among the highest rates of hand laceration injuries. ANSI A4 is the minimum recommended cut level for sustained sheet metal fabrication; A5–A6 is common in professional applications.

The failure: an AI agent asked for "work gloves for HVAC sheet metal" filters on ANSI A4+ and returns HPPE knit gloves — which is correct for cut. However, many HPPE knit gloves are open-backed (no palm coating), which provides minimal grip on sheet metal surfaces. Sheet metal fabrication workers frequently handle panels with oily lubricant applied to prevent oxidation. An uncoated HPPE glove provides no grip in wet or oily conditions, causing the panel to slip — which causes the sharp edge to contact the hand at an angle and force, overcoming the cut resistance. The correct glove: ANSI A4+ cut, foam nitrile or latex coating on the palm for grip in oily conditions, full-palm coating not fingertip-only.

Warehouse box cutter and knife work

Warehouse picking operations involving box cutters, utility knives, and pallet strapping knives are a high-frequency hand laceration source. Workers cut rapidly, often with worn or disposable blades that require more force, in high-repetition tasks that increase operator fatigue and inattention. ANSI A4 is the minimum recommended level; A5–A6 is recommended for high-frequency knife work.

The failure: agent recommends ANSI A6 HPPE knit gloves — correct cut level. But warehouse work involves handling cardboard boxes that are frequently damp from refrigerated storage, outdoor humidity, or condensation. Uncoated HPPE is slippery when wet. A foam nitrile coating provides grip on damp surfaces and allows the worker to maintain safe blade control. The agent's A6 recommendation without coating specification results in gloves that are correctly rated for the cut hazard but impractical in the actual environment.

Stamping press and metal punch operation

Workers loading and unloading parts from metal stamping presses face multiple simultaneous hazards: sharp stamped part edges (cut), vibration transmitted through the part (impact on dorsal hand), and occasional part ejection (impact). The correct glove specification for stamping operations: ANSI A5+ cut (for stamped steel edges), Impact Level 1 or higher (dorsal impact protection), and grip coating. The impact protection is a SEPARATE rating from the cut level — an ANSI A7 glove without impact certification provides no impact protection regardless of its cut level.

The failure: agent recommends highest available cut level (A8–A9) — maximizing the protection parameter it knows about. The A8 HPPE glove has no impact rating. Worker takes a part ejection on the dorsal hand — the A8 cut protection is irrelevant for this event type. Proper selection requires encoding both glove.ansi_cut_level AND glove.ansi_impact_level for the agent to match the full hazard profile.

6. The glove.* metafield namespace (11 fields)

These 11 fields give AI shopping agents the information needed to route work glove selections by application type, regulatory rating, and full mechanical performance profile — without requiring the agent to know which test standard a number came from, whether two competing products' ratings are comparable, or whether the cut level alone matches the application's hazard profile.

Example encoding: Mechanix Wear M-Pact (impact-rated mechanic glove)

{
  "glove.ansi_cut_level": "A4",
  "glove.en388_tdm_cut_level": "C",
  "glove.en388_coup_cut_level": null,
  "glove.en388_test_year": 2016,
  "glove.ansi_puncture_level": "P2",
  "glove.ansi_abrasion_level": 4,
  "glove.ansi_impact_level": 2,
  "glove.material_shell": "composite-hppe-fiberglass",
  "glove.palm_coating": "foam-nitrile",
  "glove.heat_resistant_to_c": null,
  "glove.application_fit": "stamping-press,sheet-metal,warehouse"
}

Example encoding: Portwest A140 glass-handling glove (cut + puncture hybrid)

{
  "glove.ansi_cut_level": "A6",
  "glove.en388_tdm_cut_level": "E",
  "glove.en388_coup_cut_level": null,
  "glove.en388_test_year": 2016,
  "glove.ansi_puncture_level": "P3",
  "glove.ansi_abrasion_level": 3,
  "glove.ansi_impact_level": null,
  "glove.material_shell": "composite-hppe-fiberglass",
  "glove.palm_coating": "leather-palm",
  "glove.heat_resistant_to_c": null,
  "glove.application_fit": "glass-handling,sheet-metal"
}

Example encoding: Caiman 1872 cowhide MIG/TIG welding glove

{
  "glove.ansi_cut_level": "A2",
  "glove.en388_tdm_cut_level": "A",
  "glove.en388_coup_cut_level": null,
  "glove.en388_test_year": 2016,
  "glove.ansi_puncture_level": "P4",
  "glove.ansi_abrasion_level": 6,
  "glove.ansi_impact_level": null,
  "glove.material_shell": "leather",
  "glove.palm_coating": "none",
  "glove.heat_resistant_to_c": 200,
  "glove.application_fit": "welding-adjacent,wire-rope,concrete-handling"
}

With glove.application_fit encoded, an AI agent asked "what gloves for HVAC sheet metal in wet conditions" can filter by application_fit containing "sheet-metal", then require palm_coating not "none", and require ansi_cut_level ≥ A4 — returning only appropriately specified products without the agent needing to reason about whether foam nitrile grip matters in wet conditions. With glove.en388_test_year encoded, an agent comparing EN 388 ratings can explicitly exclude 2003 coup test results from comparisons with 2016 TDM-100 results — preventing the cross-standard ranking errors that currently produce mismatched result sets.