Guide

AI in Insurance and Financial Services

Sector rules for AI in underwriting, claims, utilization review and credit decisions in Texas.

Law checked through

Short Answer

Texas insurance AI is subject to sector rules. Insurance Code § 4201.156 bars wholly or partly automated adverse utilization-review determinations, subject to its exceptions. TDI Bulletin B-0003-26, issued June 12, 2026, states expectations for AI-supported insurance decisions under existing law and human review of consequential decisions. A bulletin is agency guidance, not a separate enactment. TRAIGA preserves TDI’s insurance authority and has specified discrimination exclusions. Credit decisions require a separate ECOA, Regulation B and FCRA analysis; notice requirements differ for consumer and business credit.

Which Laws Apply

Texas AI-specific: Insurance Code § 4201.002, Insurance Code § 4201.156 and Insurance Code § 4201.303 (SB 815); Business and Commerce Code § 552.002, Business and Commerce Code § 552.056(d) and (e); Business and Commerce Code § 503.001(e) (financial institution voiceprints).

Generally applicable Texas law: Insurance Code chapters 541 (unfair practices), 542 (claims), 544 (discrimination) and 4101 (adjusters); TDI Bulletin B-0003-26; Finance Code for state-licensed lenders.

Federal: ECOA and Regulation B; FCRA; Gramm-Leach-Bliley Act.

Utilization Review (SB 815)

Insurance Code § 4201.156(a) prohibits a utilization review agent from using an automated decision system to make, wholly or partly, an adverse determination. Subsection (b) authorizes Commissioner audits and inspections; subsection (c) excepts administrative-support and fraud-detection uses. The enacted text controls over a bill analysis suggesting only a sole-basis prohibition. SB 815 took effect September 1, 2025 and applies to utilization review for plans delivered, issued for delivery or renewed on or after January 1, 2026. Insurance Code § 4201.303 includes notice-content amendments.

The Department’s 2026 Bulletin

Bulletin B-0003-26 (June 12, 2026), states that AI-made or AI-supported decisions must comply with all applicable insurance laws, may not be inaccurate, arbitrary, capricious or unfairly discriminatory, and, for consequential decisions such as claim denials, estimate reductions, fraud flags, payment delays and coverage determinations, should be reviewed and agreed to by a person before action is taken. It expects governance, testing and audit functions and notes adjuster licensing requirements.

TRAIGA and Insurers

TRAIGA does not authorize any agency other than the Department of Insurance to regulate the business of insurance (Business and Commerce Code § 552.002), and its discrimination section does not apply to insurance entities providing insurance services when they are subject to insurance law on unfair discrimination and deceptive practices (Business and Commerce Code § 552.056(d)).

Credit

Regulation B, 12 C.F.R. § 1002.9, requires the applicable adverse-action notice and either specific reasons or the prescribed notice of the right to request them, with distinct business-credit rules. A complex model does not remove the requirement applicable to the creditor. The CFPB’s May 12, 2025 withdrawal of 67 guidance documents (90 FR 20084) does not amend that regulation. FCRA may add duties when a consumer report contributes to the decision. TRAIGA Business and Commerce Code § 552.056(e) treats a federally insured institution complying with federal and state banking law as compliant with that section, not every AI-related law.

Illustrative Example (Hypothetical)

Hypothetical: a Texas lender uses an AI score in a small-business credit decision. It must determine which business-credit provision of 1002.9 applies, including the relevant gross-revenue and notice conditions, then identify the actual principal reasons supporting the decision when required. If a covered consumer report is involved, FCRA adds its own notice analysis. A generic model-score explanation does not answer each of those questions.

What Is Unsettled

How the Department will audit AI under SB 815 and the bulletin; how “wholly or partly” in SB 815 will be read.

Sources

Related Reading