01
Inconsistent input
“Consulting,” “software services,” and “IT solutions” can describe very different establishments. Keyword matching is not enough.
Turn a free-text business description into ranked U.S. SIC and NAICS codes—with complete hierarchy, calibrated confidence, alternatives, and source metadata in structured JSON.
Classify business activities—not just keywords.
Do not enter confidential or personal information in this public demo.
This preview demonstrates the response structure; it is not connected to the production classifier.
SIC 1987
6411
Insurance Agents, Brokers, and Service
NAICS 2022
524210
Insurance Agencies and Brokerages
Confidence
High · 0.97
H · Finance, Insurance, and Real Estate
64 · Insurance Agents, Brokers, and Service
52 · Finance and Insurance
5242 · Agencies, Brokerages, and Other Insurance Related Activities
Applicants and brokers describe the same business in completely different ways. Facio interprets the activity behind the words, compares it with versioned taxonomies, and returns structured data your workflow can use immediately.
01
“Consulting,” “software services,” and “IT solutions” can describe very different establishments. Keyword matching is not enough.
02
SIC and NAICS are different systems with versioned, often many-to-many relationships. Facio preserves the distinction.
03
A single code hides ambiguity. Confidence and alternatives show what can automate—and what needs review.
Use one REST endpoint to transform unstructured business activity into normalized industry data, complete hierarchy, confidence, and traceable source versions.
Send the language your customer already uses. Facio identifies the primary activity, business model, products, services, and qualifiers.
Receive the selected code and every parent level—from SIC division to industry and from NAICS sector to national industry.
When more than one classification is plausible, see the leading candidates and the distinction between them.
Use confidence bands and review reason codes to govern straight-through processing and referrals.
Every response records the taxonomy, crosswalk, classifier, and configuration versions that produced it.
Facio selects from governed taxonomy records and asks for more context when the description is insufficient.
Separate the primary business activity from products, customer type, business model, and secondary activities.
Search versioned SIC and NAICS records using lexical and semantic evidence, then compare the strongest candidates.
Return the best match, hierarchy, alternatives, confidence, warnings, and sources—or an explicit request for context.
A classification is only useful when your system understands why it was selected, where it sits in the hierarchy, and how certain the result is.
response.fields
Move from unstructured intake to consistent data without losing the original description, uncertainty, or audit trail.
Normalize the applicant’s own description before the risk reaches an underwriter.
Trigger questions, route submissions, or apply appetite rules with confidence thresholds.
Standardize legacy free-text records for consistent reporting and integrations.
Enrich onboarding data while preserving source text and classification provenance.
Reclassify historical descriptions through a reproducible, versioned process.
Aggregate exposure by SIC division, NAICS sector, or leaf industry.
Facio does not force every input into a code. It distinguishes activity from occupation, exposes one-to-many relationships, and returns an explicit review state when evidence is incomplete.
“Project manager”
An occupation is not enough to classify an establishment. What products or services does the business provide?
A predictable REST interface with JSON schemas, idempotent requests, versioned endpoints, and operational metadata.
Predictable request and response contracts for modern application stacks.
OpenAPI definitions, typed examples, and clear error responses.
API keys or OAuth, tenant quotas, and rate-limit headers.
Reproduce a result with the returned taxonomy and classifier versions.
{
"status": "classified",
"primary": {
"sic": {
"code": "6411",
"title": "Insurance Agents, Brokers, and Service"
},
"naics": {
"code": "524210",
"title": "Insurance Agencies and Brokerages"
},
"confidence": {
"score": 0.97,
"band": "high",
"decision": "auto_accept"
},
"relationship": "many_to_one"
},
"review": { "required": false },
"alternatives": [],
"versions": {
"classifier": "classifier_1.0.0",
"US_SIC_1987": "sic_1987_facio_2026_08_001",
"US_NAICS_2022": "naics_2022_facio_2026_08_001"
}
}U.S. SIC remains common in insurance, finance, and legacy datasets. NAICS is the current North American statistical standard. Facio classifies both systems independently and returns relationship provenance.
| System | SIC | NAICS |
|---|---|---|
| Facio taxonomy ID | US_SIC_1987 | US_NAICS_2022 |
| Leaf code | 4 digits | 6 digits |
| Top level | Division | Sector |
| Typical role | Legacy and industry workflows | Current U.S. statistical classification |
| Relationship | Often one-to-many or many-to-many | Often one-to-many or many-to-many |
It converts a free-text description of business activity into ranked U.S. SIC and NAICS classifications, including hierarchy, confidence, alternatives, review status, and version metadata.
Yes. Send the description your applicant, broker, or customer already uses. The API interprets the primary activity and compares it with governed SIC records instead of requiring an exact manual keyword.
Yes, when the category describes what the establishment primarily does. Details such as products, services, and customer type can improve the result.
Yes. Facio can return both U.S. SIC 1987 and U.S. NAICS 2022, including their separate hierarchies and relationship provenance.
Not necessarily. The systems were designed differently, and relationships can be one-to-many or many-to-many. Facio preserves that ambiguity.
The API can return ranked alternatives, lower the confidence band, request more context, or mark the result for human review.
No. The workflow selects from versioned taxonomy records and validates identifiers before returning them.
Yes. Response fields support intake normalization, appetite routing, referral rules, portfolio reporting, and human review.
The initial release is designed for English and Hebrew input. Additional languages can be introduced as versioned capabilities after evaluation.
Add governed SIC and NAICS classification to intake, underwriting, onboarding, and portfolio workflows through one API.
Tell us about your workflow, expected volume, and the taxonomies you need.