MCC Classification

Category assignment
you can defend.

A merchant category code decides interchange, scheme obligations, and how a portfolio looks to a regulator. Getting one wrong is expensive and quiet. Neural AI Pay treats classification as an evidence problem: several independent signals, weighed against each other, returned with the reasoning attached.

INPUTINDEPENDENT SIGNALSWEIGHED TOGETHERRESULTMANDATESMerchantdescriptorsubmitted codeBusiness registryofficial record, where availableKnown patternscurated APAC brand knowledgeSemantic matchfull published taxonomiesCorroborationagreement raises confidencedisagreement is surfacedScored resultsuggested code + confidenceconflict, where it existsranked alternativesrisk & obligation flagsApplicable mandatesscheme & regulator rulesthat govern the codefilterable by scheme,region, and country
Signals are gathered in parallel and none is decisive alone. A registry match and a text match that agree on the same code outrank either on its own — and where they disagree, the result says so rather than quietly picking a side.

Independent signals, not one guess.

Each source can reach a conclusion on its own. Agreement between sources raises confidence; disagreement is surfaced rather than averaged away. A single weak signal never carries a result by itself.

Business registry

Where a registered business number is supplied, we corroborate the merchant against the official registry record rather than trusting the descriptor alone.

Known patterns

A curated body of merchant-naming patterns and regional brand knowledge, maintained by us — the part that understands what an APAC trading name actually implies.

Semantic match

The descriptor is matched against the full published category taxonomies for both Visa and Mastercard, so wording that no rule anticipates still lands somewhere defensible.

What comes back.

A score with no reasoning is not much use in a compliance file. Every result carries the basis for the decision.

Suggested code
The best-supported category for the merchant, with a confidence score.
Conflict flag
Raised when the submitted code looks unlikely and a better-supported one exists.
Alternatives
Ranked runners-up with the reasoning behind each, so a reviewer can judge rather than defer.
Risk & obligations
Scheme restrictions, risk classification, and Enhanced Due Diligence requirements attached to the code.
Applicable mandates
Scheme and regulator mandates that govern the resulting code, filterable by scheme, region, and country.
Audit record
Every validation logged and exportable as evidence for an auditor or regulator.

Kept current against the published rules.

The category data behind the engine is checked against the category standards Visa and Mastercard publish, and every code we hold is traceable to at least one of them. When a scheme issues a new edition, the differences are reviewed before anything changes — no code enters the system on an automated guess.

Reviewed, not scraped

Risk classification, scheme restrictions, and Enhanced Due Diligence requirements are not published by any scheme in a form that can simply be imported. They are our own compliance judgement, reviewed deliberately and revisited as rules move — which is the part of this that a scraper cannot reproduce.

Try it on a real descriptor.

Five free checks, no account required.

Try MCC ClassificationAPI documentation →