Give us a business name and address — Classification returns a full multi-taxonomy profile across NAICS, SIC, workers' comp and ISO general liability, each code carrying its own confidence and a reasoning trail you can defend.
Underwriting, rating and audit each speak a different code language. Classification resolves a single business to all of them at once — cross-walked, scored, and explainable — so every downstream system gets exactly the taxonomy it expects.
NAICS, SIC, NCCI & state workers'-comp, and ISO GL resolved together and cross-walked from a single lookup — not four separate tools.
Each classification carries its own calibrated confidence, so a firm match and a best guess are never treated as the same thing.
Trace any code back to the live signals that produced it — the transparent trail both underwriting and audit need.
A misclassified business rates on the wrong table, lands outside the right appetite, and surfaces again at audit as premium leakage. Done by hand, classification is slow, inconsistent between analysts, and impossible to explain months later.
Illustrative result. One lookup returns every code the business touches, cross-walked and individually scored.
A NAICS lookup matches words. NeuralMetrics reads the business — its operations, hazards and context — and reasons its way to the right code, then extends it two digits further into the sub-classes underwriting actually prices on.
Most classifiers stop at a single NAICS code. NeuralMetrics resolves the full profile across the taxonomies that actually drive underwriting, rating and audit — and keeps them aligned to one another.
2017 and 2022 NAICS mapped in parallel at 2-, 6- and 8-digit depth, so codes line up with whatever your rating and reference tables expect.
NCCI class codes plus the independent-state bureaus that don’t follow them, so the WC class matches the jurisdiction it will actually be written in.
ISO general-liability and SIC out of the box — plus a bridge to your own in-house code set, mapped from the same resolved profile.
Submissions rarely arrive clean. Classification is engineered to find the business behind an imperfect name, a suite-less address or a legal entity that shares nothing with the trading name.
Resolves DBAs, abbreviations, misspellings and legal-vs-trading-name mismatches to the same underlying business.
Uses the location as a second signal — so a common name at a specific address still resolves to the right operation.
Draws on current signals about what the business actually does, rather than a single static reference file that ages the day it ships.
A name and address is enough — one record, or a whole book in a single batch.
The business is identified, verified against live evidence, and mapped across every taxonomy.
Every code comes back with its confidence and the signals behind it — ready for rating, appetite and audit.
One classification call returns every code your workflow needs — and it isn’t US-only. The same structured government sourcing runs across 35+ countries and 22+ languages.
The standard six-digit code, extended two digits into underwriting-level sub-classifications.
Full Standard Industrial Classification coverage for legacy and rating systems.
General-liability and WC class codes aligned to the operations that drive them.
Your proprietary classification scheme, mapped to NeuralMetrics' expanded index.
Classification is where the platform starts. Accurate class codes feed appetite matching and eligibility — so the decision inherits the right answer instead of the wrong one.
Bring a handful of tricky risks — we’ll show the codes, the sub-classes and the sources behind each one.
Request a demo