What is AIAG damage grading?

AIAG damage grading is a standardized way to describe vehicle damage developed in the automotive supply chain to keep condition assessments consistent across inspectors and locations. Instead of free-text notes like 'small dent,' it uses defined severity levels so the same damage gets the same grade no matter who's looking — which is what makes condition reports comparable and defensible.

What does the K1–K5 scale mean?

K1–K5 is a five-level severity scale for vehicle damage, running from K1 (minor, cosmetic) up to K5 (severe, structural or functional). Lower numbers are light wear that may be acceptable; higher numbers are significant damage that typically requires repair and recharge. Putting a number on severity is what turns a subjective 'is this bad?' into a consistent, comparable grade.

How does AI apply a consistent damage grade?

An AI inspection model is trained on graded damage examples, so it maps a photo to a severity level the same way every time — removing the inspector-to-inspector drift that plagues manual grading. The result comes back as a structured grade alongside the pass/fail criteria, so every vehicle is assessed against the same standard.

What is the difference between a K2 and a K3 damage grade?

K2 is light damage at the cosmetic end: visible but small, such as a shallow scratch in the clear coat or a minor ding with no paint break. K3 is moderate damage that clearly needs attention, such as a dent in a panel, a scratch down to primer, or cracked trim. The practical difference is disposition: a K2 may or may not be chargeable under a given policy, while a K3 is generally repairable and commonly chargeable.

Who defines the K1 to K5 thresholds?

The AIAG (Automotive Industry Action Group) is an automotive industry association whose standardization work underpins condition-reporting conventions across the supply chain, and K-style severity scales are how that standardization shows up in vehicle inspections. Exact thresholds are set by the individual program, lessor, or fleet policy that adopts the scale, so a K3 in one damage matrix may sit slightly differently in another. What stays constant is the shape: five defined levels running from cosmetic to severe.

Does a damage grade include a repair cost estimate?

No. A grade describes severity, not price. It tells you how bad the damage is on a defined scale so that two inspections of the same panel land in the same place, but it does not price parts, paint, or labor. VerifyAI documents and grades damage and exports a condition report; it does not currently generate repair-cost estimates. Teams that need appraisal-grade cost figures pair grading with an estimation tool or their own rate matrix.

AIAG & K1–K5 Damage Grading Explained (With Examples)

AIAG damage grading is a standardized way to describe vehicle damage, and the K1–K5 scale is the five-level severity ladder it is usually expressed with: K1 is minor cosmetic wear, K2 is light damage, K3 is moderate damage that clearly needs attention, K4 is significant damage to a panel or component, and K5 is severe structural or functional damage. The point of a grade is comparability — the same damage gets the same number regardless of who is looking, which is what makes a condition report auditable and defensible in a dispute.

If you've read a condition report and seen a damage entry tagged "K2" or "K4," that's severity grading. Standardized grades exist because free-text notes don't survive scrutiny: "small scratch" means something different to everyone, and "K1" doesn't. What follows is the scale grade by grade with concrete examples, and how an AI inspection model applies the same grade every time.

In brief

The K1–K5 scale runs minor to severe: K1 cosmetic, K2 light, K3 moderate, K4 significant, K5 severe.
A grade describes severity, not repair cost. It standardizes the assessment; pricing is a separate step.
Exact thresholds are set by the adopting program or fleet policy, so grades are most comparable within a program — the five-level shape is what's constant.
Grading only changes outcomes when it is applied consistently. Manual grading drifts between inspectors and across a shift; a model trained on graded examples does not.
The strongest damage record is a grade plus a before/after delta — severity, and proof the damage is new.
VerifyAI returns the grade as a structured field next to the per-criterion pass/fail result, so downstream code can branch on it.

A note on positioning: VerifyAI leads with parking compliance, and vehicle damage inspection is a secondary capability. We publish this explainer because grading is foundational to defensible condition records, and the topic is poorly explained elsewhere.

Vehicle appraisal pricing: $2.00 per completed appraisal, including up to 16 original photos and detailed damage analysis. Single-photo API checks are a separate product and do not include the appraisal workflow. See the offer and arrange production access.

Why standardized damage grading exists

Standardized grading exists because free-text damage notes can't be compared, audited, or defended. A defined scale replaces an opinion with a classification, so two assessments of the same panel land in the same place.

Damage assessment used to be free text. One inspector writes "minor dent," another writes "noticeable ding," a third just checks a box. When those notes feed a rental dispute, a lease return, or a remarketing listing, the inconsistency is a problem: you can't compare two reports, you can't audit a pattern, and a customer can always argue the wording was subjective.

The automotive supply chain solved this with standardized grading — defined severity levels that mean the same thing regardless of who's assessing. The AIAG (Automotive Industry Action Group) is one of the bodies behind this kind of standardization, and severity scales like K1–K5 are how the standard shows up in practice. The point is comparability: a grade is a shared vocabulary that makes condition records consistent and defensible.

Three ways of recording the same dent, and what each one is worth when someone pushes back:

Approach	Consistency between assessors	Auditable across a fleet	Strength when disputed
Free-text note ("minor dent")	None — wording is personal	No, not comparable	Weak; the customer argues the wording
Defined scale, graded by hand	Better, but drifts by inspector and shift	Partial; depends on discipline	Moderate; the scale helps, the drift hurts
Defined scale, applied by a model	Same input maps to the same grade	Yes — every asset on one standard	Strong; a classification, not an opinion

Grade ≠ opinion

The whole value of a grade is that it removes opinion. "That looks pretty bad" is an argument. "That's a K3" is a classification against a defined scale — the same input maps to the same output, which is exactly what you want when a charge is being disputed.

The K1–K5 severity scale, grade by grade

K1–K5 runs from minor to severe: K1 is cosmetic wear, K2 is light damage, K3 is moderate damage needing attention, K4 is significant panel or component damage, and K5 is severe structural or functional damage. Exact thresholds vary by program, but the shape is consistent.

Grade	What it means	Example	Typical disposition
K1	Minor / cosmetic — normal use	Faint scuff that buffs out, tiny stone chip, light swirl marks	Usually accepted as wear; no repair
K2	Light — visible but small	Shallow scratch through the clear coat, minor ding with no paint break	May or may not be chargeable, per policy
K3	Moderate — clearly visible, needs attention	Dent in a panel, scratch down to primer, cracked trim piece	Generally repairable and commonly chargeable
K4	Significant — affects a panel or component	Deep dent with creasing, cracked bumper, damaged glass	Repair or replacement required
K5	Severe — structural or functional	Crushed panel, shattered windshield, missing or destroyed component	Major repair; clear high-severity charge

The same levels in prose, with a little more nuance on each:

K1 — Minor / cosmetic. Light, surface-level wear that's often acceptable as normal use. Examples: a faint scuff that buffs out, a tiny stone chip, light swirl marks. Usually no repair needed.
K2 — Light. Visible but small damage at the cosmetic end. Examples: a small shallow scratch through the clear coat, a minor ding with no paint break. May or may not be chargeable depending on policy.
K3 — Moderate. Clearly visible damage that typically needs attention. Examples: a dent on a panel, a scratch down to primer, a cracked trim piece. Generally repairable and commonly chargeable.
K4 — Significant. Substantial damage affecting a panel or component. Examples: a deep dent with creasing, a cracked bumper, damaged glass. Requires repair or replacement.
K5 — Severe. Major structural or functional damage. Examples: a crushed panel, a shattered windshield, a missing or destroyed component. Significant repair, and a clear, high-severity charge.

The progression is intuitive — bigger number, worse damage — but the value is that "bigger number" is defined, not guessed. A K3 from one assessment is comparable to a K3 from another.

How AI applies a consistent grade

An AI inspection model applies a consistent grade by mapping the photo to a severity level with the same trained criteria on every call, then returning that level as a structured field rather than a note. There is no shift drift and no inspector-to-inspector disagreement to reconcile later.

Manual grading still drifts. Even with a defined scale, a tired inspector at the end of a shift grades differently than a fresh one at the start, and two people will disagree on whether a given dent is a K2 or a K3. That inconsistency is exactly what undermines a report when it's challenged.

An AI inspection model is trained on graded damage examples, so it maps a photo to a severity level the same way every time — no shift drift, no inspector-to-inspector disagreement. In practice, VerifyAI returns the grade as a structured field alongside the pass/fail criteria:

The model detects damage against the policy criteria (no body damage, no broken glass, no missing parts).
It assigns a severity grade on the AIAG / K1–K5 scale.
The grade comes back in the response, so every vehicle is assessed against the same standard automatically.

For the implementation detail, see the AIAG and K-grades guide, and for how grading fits the full inspection, the vehicle damage inspection use case. The methodology behind how VerifyAI assigns grades is on the methodology page. If you want quick definitions, the glossary covers the AIAG damage grade and the K1–K5 scale. The criteria a grade is evaluated alongside are configured as policy-as-code, so the same standard applies at every site.

Why the grade matters in a dispute

A grade matters in a dispute because it converts a judgement call into a classification the other side has to argue with on its own terms. It does that in two ways:

Consistency across the fleet. Every vehicle graded the same way means you can audit, compare, and trust your own data. A pattern of K4s at one location is a signal you can actually act on.
Defensibility in a chargeback. A consistent severity grade, paired with a before/after delta that isolates new damage from pre-existing wear, turns a subjective argument into a documented classification. "It's a K3 that wasn't there at pickup" is far stronger than "it looks damaged to us." Packaged as a PDF condition report, the grade, the baseline, and the delta travel together into a representment.

Where the grade actually gets used, and what has to sit next to it to be worth anything:

Moment	What the grade decides	What it must be paired with
Rental or fleet check-out	The baseline condition on record	A customer-acknowledged capture — see self-inspection links
Rental or fleet return	Whether damage is chargeable and at what severity	A before/after delta isolating new damage
Lease return or remarketing	How the unit is described to the next buyer	A consistent grade applied across the whole fleet
Chargeback representment	The severity claim the charge rests on	Timestamped capture plus an exported condition report
Internal audit	Where damage clusters by site, class, or driver	The same grading standard applied everywhere, every time

On compliance claims

VerifyAI is GDPR-aligned, with a SOC 2 audit in progress — not yet SOC 2 certified. See security and GDPR for the current status.

See grading on a real photo

The fastest way to understand the scale is to grade a real photo and read the result.

Start free in the sandbox — $5 in credit, no card required. Upload a damage photo and see the structured grade come back. To wire it into your own pipeline, see the damage inspection API quickstart; for the workflow around a return, rental return verification; for cost, vehicle-appraisal pricing is $2.00 per completed appraisal with up to 16 photos; and for a survey of inspection tools, the best vehicle damage detection software roundup.

A standardized grade is a small thing that fixes a big problem: it makes every condition assessment mean the same thing — to your team, to your data, and to a card issuer reviewing a dispute.