AI Disclosures

Every output produced by Stratifier — including strategy documents, marketing plans, competitive research, assessments, dashboards, and chat responses — is advisory decision support. Outputs must not be used as the sole basis for any decision about an individual person, including but not limited to hiring, firing, promotion, compensation, discipline, benefits, credit, housing, education, healthcare, insurance, immigration, or legal matters.

A qualified human must review, edit, and take responsibility for any action taken based on Stratifier output. Stratifier does not make automated decisions about individuals.

Wherever AI-generated content appears in Stratifier, a visible label reads "AI-generated — verify before use." The label links back to this page. When content touches HR or people topics, we add an extra reminder that it is not for individual employment decisions.

Every PDF and Word document Stratifier generates includes both a visible disclosure (cover page and footer) and machine-readable provenance metadata. That metadata includes:

• A flag that the document is AI-generated with human review.
• The model family and provider that produced the draft.
• The date the draft was generated.
• A SHA-256 hash of the generated content body.

The hash is also recorded in Stratifier's internal provenance log so anyone can verify a document's origin using our public detection endpoint.

Anyone — including recipients of a Stratifier document — can check whether a piece of content was generated by Stratifier at /ai-detection. The endpoint accepts either a SHA-256 content hash or a Stratifier PDF and returns the model family, provider, and generation date associated with the content. No personal data or company identifiers are returned.

A public summary of every AI capability in Stratifier — what it does, what data it reads, which model family produces it, the human-review checkpoint, and known limitations — is published at /ai-systems. The inventory is reviewed at least annually and after material changes.

Stratifier processes company data through the Lovable AI Gateway. Provider training on Stratifier prompts and outputs is disabled by default under the Gateway's terms. Optional integrations (Semrush, Google Analytics, Google Search Console) run only when a company connects them and only against that company's own data.

Stratifier logs the surface, model, and content hash of each generation for audit purposes; it does not retain full prompt or output text in the provenance log. Full reports are saved only where the customer has explicitly saved them into their workspace, and are deleted with the account through Settings.

If a Stratifier output is inaccurate, misleading, biased, or otherwise concerning, report it from any AI panel using the "Report an issue" link, or email support@nextrockadvisors.com. We review reports weekly and log them internally for our AI risk management program.

Every company on Stratifier can opt into the SB 942 compliance path from its Company Profile. When enrolled, the company gets a one-time AI-terms consent recorded per terms version, a persistent on-screen "AI-generated" badge on every AI panel, provenance rows for each generation, and provenance metadata embedded in every exported PDF and Word document. Consultants can monitor enrollment, consent status, provenance activity and open incidents across their book of business at /app/consultant/compliance; master admins get the same view plus the incident workqueue under AI Review.

These practices are designed to meet or exceed the California AI Transparency Act (SB 942), the disclosure and risk-management expectations of the Colorado AI Act (effective 2026), and the NIST AI Risk Management Framework (Govern, Map, Measure, Manage). Because Stratifier outputs are advisory only, Stratifier is not a "high-risk" system under the Colorado AI Act.