Texas Drew the Line on AI. Here’s How Employers Are Walking It.

Across the country right now, institutions are grappling with a deceptively simple question: What do we do with artificial intelligence? Universities are split on whether to allow or restrict it. The debate is playing out in boardrooms, on factory floors and inside HR departments with equal intensity. But in Texas, the conversation has moved past the existential and into the operational.

That should surprise no one. Texas is a business metropolis — Austin’s tech corridor is home to Tesla, Oracle, Samsung and a growing concentration of AI companies; Dallas-Fort Worth houses one of the largest clusters of Fortune 500 headquarters in the country and a dominant financial services sector; and Houston, one of the most diverse cities in the nation, anchors the global energy industry and the Texas Medical Center. When a state with that collective economic footprint decides to build a regulatory framework around AI, it is worth paying attention.

We are living through a technological inflection point no less significant than the rise of the internet or the digitization of commerce. Gone are the days when AI was speculative hype or a chatbot novelty. It has rolled out. It is embedded in operations. It is changing how we do everything — how we hire, how we evaluate performance, how we manage risk. And now, it is up to employers to take the bull by the horns: shape the infrastructure, integrate it responsibly and govern its use. Nobody wants to be left in the proverbial dust.

Over six months ago, when the Texas Responsible Artificial Intelligence Governance Act (TRAIGA) took effect Jan. 1, much of the legal commentary focused on what the law says. Now it’s time to talk about what employers are actually doing with it.

For general counsel and HR leaders at Texas companies, TRAIGA isn’t an abstract compliance exercise — it’s the backdrop against which real decisions are being made every day about résumé screening algorithms, AI-assisted performance reviews, predictive scheduling tools and workforce analytics platforms. The question is no longer whether AI belongs in the workplace. It’s how to use it without creating liability.

AI in the workplace is not going away — that is the reality for general counsel and HR leaders. It is a fundamental shift in how businesses operate, and the companies that thrive will be those that embrace it thoughtfully, with eyes open to both the opportunity and the risk. But here is what experience has made clear: AI is only as effective as its user is strategic and only as safe as its user is cautious. As with most things in life, what you put in is what you get out. The quality of governance determines the quality of outcomes.

Texas is leading the way. While states like Georgia are taking a piecemeal, sector-by-sector approach to AI regulation, and New York City has zeroed in specifically on hiring — requiring annual independent bias audits of automated employment decision tools under Local Law 144 — Texas chose to regulate broadly, establishing a single omnibus framework that applies across industries and use cases. Other states will follow, one way or another. The employers who get ahead of the curve now, under TRAIGA’s framework, will find themselves better positioned regardless of what other jurisdictions do next.

This article maps TRAIGA’s framework onto the employer use cases where the rubber meets the road, offering a practical guide for companies that are deploying, or considering deploying, AI tools across the employment lifecycle.

The Employer’s Paradox: Covered but Carved Out

The first question any Texas employer asks: Does TRAIGA apply to my workforce AI tools?

The short answer is yes — but not in the way you might expect. TRAIGA covers any entity doing business in Texas or deploying AI in the state, and its definition of “artificial intelligence system” is broad enough to reach everything from a large language model to a predictive maintenance algorithm on a plant floor.

Here is where it gets nuanced. The law’s definition of “consumer” excludes individuals acting in an employment context — so TRAIGA’s disclosure obligations do not apply to employer-employee interactions. Many employers have read that carveout and stopped there. That is a mistake. The statute’s prohibition on intentional discrimination applies to any “person,” in any context, across the entire life cycle of an AI system — including hiring, performance management and workforce planning. The disclosure carveout does not shield employers from the antidiscrimination mandate.

Use Case 1: AI-Powered Hiring and Applicant Screening

Perhaps no employer use case draws more scrutiny than AI in recruitment. Applicant tracking systems that use machine learning to rank résumés, chatbot-driven initial interviews and predictive models that score candidates on “culture fit” are now commonplace at major Texas employers — and all fall squarely within TRAIGA’s definition of an AI system.

The critical distinction under TRAIGA is one that should relieve (but not relax) employers: The statute requires intent to discriminate. Unequal outcomes alone do not establish a violation. An AI hiring tool that produces statistically skewed results — more men than women advancing to final-round interviews, for example — does not violate TRAIGA if it was not designed with discriminatory purpose.

But here is what in-house counsel should not lose sight of: Federal law tells a different story. Title VII of the Civil Rights Act of 1964, the Age Discrimination in Employment Act and the Americans with Disabilities Act all permit claims based on disparate impact, no intent required. So while TRAIGA provides Texas-specific breathing room, employers deploying AI hiring tools remain fully exposed under federal frameworks. The practical implication is that companies cannot treat TRAIGA compliance as a ceiling. It is a floor.

What to do now: Document the legitimate business purpose behind each AI hiring tool. Ensure that vendors can articulate (and evidence) the nondiscriminatory design philosophy underpinning their algorithms. Maintain records showing that AI screening outputs are reviewed by human decision-makers before adverse actions are taken.

Use Case 2: Performance Management and Workforce Analytics

AI-driven performance management is the use case many employers don’t realize they already have. If your organization uses tools that aggregate productivity data, flag underperformance through algorithmic scoring or recommend personnel actions based on predictive models, you are deploying AI in a context that TRAIGA’s discrimination prohibition reaches.

Consider a distribution center that uses AI to monitor warehouse workers’ pick rates, identify “efficiency outliers” and generate termination recommendations. Or a professional services firm that uses natural language processing to evaluate the quality of written work product and recommend promotion candidates. These tools generate outputs that influence employment decisions — decisions that touch on every protected class that TRAIGA enumerates.

The safeguard here is governance. TRAIGA rewards companies that maintain documented AI use policies covering ownership, escalation paths and monitoring protocols. And companies that substantially comply with the NIST AI Risk Management Framework have a statutory defense against enforcement under Texas Business & Commerce Code Section 552.104.

What to do now: Create an internal AI inventory that captures every tool touching performance evaluation or personnel actions. For each tool, document: what data it ingests, what outputs it produces, who acts on those outputs and what human review occurs between algorithmic recommendation and employment decision.

Use Case 3: Biometric AI on the Plant Floor

For manufacturers and industrial employers — a significant constituency in Texas — AI systems that process biometric data present a distinct set of considerations under TRAIGA. Facial recognition for facility access, gait analysis for workplace safety monitoring and voice authentication systems all fall within this category.

TRAIGA provides a meaningful exemption here: The training and development of AI models is generally carved out from the consent-and-retention requirements of the Texas Capture or Use of Biometric Identifier Act (CUBI). But the exemption has hard limits. If the system moves beyond training a model in the abstract and is used to uniquely identify a specific individual, or if biometric data collected for training purposes is later repurposed for a different commercial use, full CUBI compliance, including its penalty provisions under Section 503.001, snaps back into place.

The practical risk for manufacturers is mission creep. A facial recognition system initially deployed for access control that later gets repurposed to monitor employee attendance or identify workers not wearing PPE may cross the line from exempt training use to nonexempt identification — triggering consent obligations the employer never anticipated.

What to do now: Maintain clear written documentation of the stated purpose for each biometric AI system. Build internal controls that require legal review before any biometric tool is repurposed or expanded beyond its original scope.

Use Case 4: Predictive Scheduling and Wage Optimization

AI-powered scheduling tools are now standard at Texas retailers, restaurants, healthcare systems and logistics companies. These platforms use machine learning to forecast demand, optimize labor allocation and minimize overtime costs. Under TRAIGA, they qualify as AI systems if they generate outputs — such as scheduling recommendations and staffing predictions — that influence the work environment.

While these tools don’t raise the same discrimination concerns as hiring algorithms, they intersect with TRAIGA in a less obvious way: the attorney general’s investigative authority. Under Section 552.103, the attorney general can demand documentation on an AI system’s purpose, training data, inputs, outputs, performance metrics, known limitations and post-deployment safeguards. A scheduling tool that systematically assigns less-desirable shifts to workers in a particular demographic could, if the pattern is sufficiently stark and other evidence supports intent, trigger an inquiry.

More immediately, workforce-scheduling AI intersects with federal wage-and-hour obligations. An algorithm that shaves minutes from shifts, rounds time entries or fragments full-time schedules into part-time blocks may generate Fair Labor Standards Act or state wage-and-hour exposure that no state AI law immunizes.

What to do now: Ensure scheduling AI vendors can produce the documentation the attorney general’s office may request on short notice. Periodically audit scheduling outputs for demographic patterns that, even if not violating TRAIGA, could invite scrutiny.

The Vendor Relationship Is the Compliance Gap

Across all these use cases, a single theme emerges: The employer’s compliance posture is only as strong as its vendor relationships. Most Texas employers do not build their own AI tools; they buy or license them. When the attorney general exercises investigative authority, the employer will need to produce documentation about system design, training data and performance metrics that live in the vendor’s possession, not its own.

This is where procurement teams and employment lawyers need to converge. AI vendor contracts should include robust documentation obligations, cooperation clauses for regulatory inquiries, representations about nondiscriminatory design and indemnification for enforcement actions arising from system defects.

What’s Coming: The Sept. 1 Deadline

The attorney general’s online complaint portal goes live by Sept. 1 — less than two months from now. Once consumers and employees have a streamlined mechanism to submit complaints, the enforcement landscape shifts from theoretical to operational. The window for preparation is closing — and employers that have spent the first half of 2026 observing from the sidelines are running out of runway.

On the federal front, Congress attempted last year to impose a 10-year moratorium on state AI enforcement through the One Big Beautiful Bill Act — a move that would have effectively put TRAIGA on ice. The Senate killed that proposal in a 99-1 vote July 1, 2025. But preemption efforts are not dead. Congressional leaders have pledged to pursue similar legislation, and an executive order issued in late 2025 asks agencies to identify state AI laws that may impede innovation. Until preemption actually passes, if it passes, TRAIGA remains fully enforceable, the attorney general’s office is staffing up and the compliance portal clock is ticking.

The employers best positioned for what comes next are those treating this window not as a reprieve but as a head start.

A point worth making plainly: This is a fluid situation, and no one has all the answers yet. Not the attorney general’s office, not the vendors selling AI tools and not the lawyers advising on compliance. What we do know — what we have been emphasizing to our manufacturing and technology clients — is that the foundation of any defensible AI strategy is governance. A written AI governance policy is not a luxury or a future aspiration. It is the single most concrete step an employer can take right now to demonstrate reasonable care, satisfy the documentation demands the attorney general’s office will inevitably make and create the institutional muscle memory that keeps pace as the law evolves.

The landscape is evolving fast and is unforgiving of hesitation. There is understandable caution, but every major technological revolution rewards those who run toward it, not away from it. Texas employers have a framework. The question now is whether they will use this moment to lead or to lag.

The answer should be straightforward: embrace the technology, build the guardrails as you go and recognize that the willingness to move forward — imperfectly, iteratively, but deliberately — is itself the competitive advantage.