Selling an AI-Powered Law Practice in Texas: Ethics, Valuation, and Deal Terms for the 2026 Succession Wave

Why AI Changes the Practice Sale Equation

The large cohort of retirement-age solo and small-firm Texas attorneys has created an active market for practice transitions. But the traditional guides to selling a law practice — which focus on goodwill valuation, client notice procedures, and professional liability insurance — were written for a world where the practice's technology stack was a filing cabinet and a case management subscription. In 2026, that is no longer the world most Texas attorneys occupy.

If your practice has integrated generative AI tools — for legal research, document drafting, contract review, or client communication — the sale of that practice now involves assets and liabilities that conventional deal frameworks do not address. AI vendor contracts carry data processing terms that may not be assignable. Client confidential information may reside inside AI platforms in ways that complicate file migration. And the Texas Responsible Artificial Intelligence Governance Act (TRAIGA), which took effect January 1, 2026, has made compliance documentation a component of the practice's value — and its risk profile.

We have written previously about the ethics requirements for selling a law practice under Model Rule 1.17 and about why succession is fundamentally about client continuity, not just a transaction. This article addresses the layer that those guides do not cover: what happens when the practice being sold has AI infrastructure embedded in its operations, and how Texas's new regulatory and ethical landscape creates deal terms that buyers and sellers must negotiate before closing.

TRAIGA Compliance as a Transferable Asset — and Liability

TRAIGA, signed into law on June 22, 2025, and effective January 1, 2026, applies to "covered persons and entities" who develop or deploy AI systems in Texas — defined broadly as "any person who promotes, advertises, or conducts business in Texas" or "develops or deploys an artificial intelligence system in Texas," according to the Greenberg Traurig analysis of the statute. A law firm that deploys AI tools for client work falls within this definition as a deployer.

The law imposes several obligations that become part of the practice's compliance posture:

• Prohibited uses: A person may not use AI to incite crime or violence, infringe constitutional rights, or unlawfully discriminate against a protected class.
• Civil penalties: Curable violations carry $10,000–$12,000 per violation; uncurable violations range from $80,000–$200,000; ongoing violations accrue at $2,000–$40,000 per day.
• Safe harbor: A person who substantially complies with the NIST AI Risk Management Framework or similar recognized standards is not liable under TRAIGA.
• State agency enforcement: If the Attorney General finds that a person licensed by a state agency has violated TRAIGA, the agency may impose sanctions including license suspension, revocation, or fines up to $100,000.

For a law practice sale, this means two things. First, the practice's TRAIGA compliance documentation — AI system inventories, risk assessments, NIST framework alignment records, vendor vetting files — is a transferable asset. A buyer who inherits a practice with documented compliance has a running start on their own obligations. Second, undocumented or non-compliant AI use is a liability that transfers with the practice. A buyer who discovers post-closing that the seller was using a consumer-grade AI tool that trained on client inputs may face both TRAIGA exposure and disciplinary risk — and the deal's indemnification provisions become the primary battleground.

Note that TRAIGA's enforcement landscape is not static. A December 2025 executive order directed the U.S. Attorney General to establish an AI Litigation Task Force to challenge state AI laws on preemption grounds. As of mid-2026, TRAIGA remains in effect, but the federal preemption fight creates uncertainty that deal documents should acknowledge — typically through a regulatory change provision that addresses what happens if TRAIGA is partially or fully preempted during the earn-out period.

AI Vendor Contracts: What Transfers and What Doesn't

The State Bar of Texas AI Toolkit provides guidance on evaluating AI tools and vendors, including sample forms for client disclosures and checklists for evaluating AI tools. It underscores that vendor vetting is now part of a lawyer's professional obligations. Opinion 705 of the Professional Ethics Committee, issued in February 2025, recommends that lawyers review the terms of service of any generative AI tool, learn about the provider's data-security protections, and train staff on appropriate use — precautions the Committee traces to its earlier Opinion 680 on cloud computing.

In a practice sale, these vendor relationships become deal items. The key questions are:

Assignment Restrictions

Most enterprise AI vendor contracts contain anti-assignment clauses requiring vendor consent for transfer. A seller cannot simply hand over a ChatGPT Enterprise or Copilot for Microsoft 365 account to the buyer. The deal timeline must include a vendor-consent workstream, and the purchase agreement should make closing contingent on key vendor assignments.

Data Processing Agreements

Enterprise AI tiers typically include data processing agreements (DPAs) that prohibit the vendor from training on user inputs, specify data retention limits, and require breach notification. These DPAs are practice assets — but they are also personal to the contracting party. The buyer must either negotiate new DPAs or obtain vendor consent to assume the existing ones. If the seller's DPA was negotiated on favorable terms (e.g., zero data retention, SOC 2 Type II certification), that has quantifiable value.

Vendor Vetting Documentation

The due diligence files that the seller compiled when evaluating AI vendors — security questionnaires, SOC reports, privacy policy reviews, and the checklist the State Bar's AI Toolkit recommends — should transfer as part of the practice's compliance infrastructure. A buyer who inherits this documentation can demonstrate that vendor selection was conducted with reasonable diligence, which matters both ethically and in any future TRAIGA enforcement action.

Subscription Economics

AI tool subscriptions are typically priced per seat or per usage. The buyer needs to understand not only the current cost but the cost trajectory — many vendors have raised enterprise pricing significantly as adoption has scaled. The deal's working capital adjustment and the buyer's post-closing budget should account for these costs.

Client Confidential Information in AI Systems: The Rule 1.05 Problem

Texas Disciplinary Rule 1.05 defines "confidential information" broadly — encompassing both privileged information protected by the attorney-client privilege and "unprivileged client information," meaning "all information relating to a client or furnished by the client, other than privileged information, acquired by the lawyer during the course of or by reason of the representation of the client." The rule prohibits a lawyer from knowingly revealing confidential information except as permitted by specified exceptions.

Opinion 705 addressed this duty directly in the AI context. The Committee warned that "self-learning" AI programs — those that store and incorporate user inputs into their training datasets — pose a risk that "confidential information a lawyer inputs to the program may be stored within the program and revealed in responses to future inquiries by third parties. That is obviously unacceptable." The opinion requires lawyers to be "reasonably satisfied that the program will not reveal confidential information to others" before inputting any client information.

This creates a unique problem for practice sales. Client files in a traditional practice are self-contained — they live in a case management system, a document repository, or physical storage. When the practice transfers, the files transfer with it. But client information that was input into a generative AI tool may not be fully exportable. The AI platform may retain prompts, generated outputs, and training data derived from those inputs on its servers, even after the seller cancels the subscription.

This raises several deal-specific issues:

Data Export and Deletion

Before closing, the seller should attempt a complete data export from each AI platform — including all prompts, generated documents, and conversation logs. The seller should also request that the vendor delete all retained data, though this may not be possible if the vendor's terms permit retention for training purposes. The purchase agreement should allocate responsibility for any data that cannot be exported or deleted.

Disclosure to Buyer

Under Rule 1.05(c), a lawyer may reveal confidential information when "expressly authorized to do so in order to carry out the representation" or when "the client consents after consultation." Providing client files to a prospective purchaser requires client consent — and the same principle applies to client information stored in AI systems. The seller must obtain client authorization before disclosing which AI tools were used and what information was input, and the deal's client notice package should address this.

Ongoing Confidentiality Risk

If the seller used a self-learning AI tool that retained client inputs, the confidential information may persist on the vendor's servers indefinitely. Neither the seller nor the buyer can fully remediate this risk. The deal documents should disclose it, and the buyer should factor it into both the purchase price and the client communication strategy post-closing.

Supervision Policies and Documentation Under Rule 5.03

Texas Disciplinary Rule 5.03 requires that a lawyer with direct supervisory authority over a nonlawyer "shall make reasonable efforts to ensure that the person's conduct is compatible with the professional obligations of the lawyer." The rule applies to "a non-lawyer employed or retained by or associated with a lawyer" — and as we have discussed in our AI ethics compliance guide, the emerging consensus treats AI tools functionally as nonlawyer assistants whose output must be supervised and verified.

A practice that has integrated AI tools should have written supervision policies: which tools are approved, what types of client information may be entered, what verification steps are required before AI-generated work product is used, and how staff are trained. The State Bar's AI Toolkit includes sample forms for client disclosures and checklists for evaluating AI tools — the kind of documentation that, when maintained and updated, demonstrates compliance with Rule 5.03's reasonable-efforts standard.

In a practice sale, these policies are both assets and obligations. The buyer inherits not just the tools but the supervisory framework — and if that framework is weak or undocumented, the buyer faces immediate compliance risk. The deal's due diligence should include a review of:

• Written AI use policies (if they exist)
• Staff training records on AI ethics
• Engagement letter disclosures regarding AI use
• Vendor vetting files and security assessments
• AI system inventories mapped to TRAIGA compliance requirements

If the seller has not maintained this documentation, the buyer should negotiate either a pre-closing remediation period or a purchase price adjustment that reflects the cost of building the compliance infrastructure from scratch.

Deal Terms Unique to AI-Powered Practices

Bringing these threads together, several deal terms appear in AI-powered practice sales that traditional practice purchase agreements do not address:

AI Compliance Representations and Warranties

The seller should represent that all AI tools used in the practice have been vetted for confidentiality and security, that all vendor contracts are in good standing, and that no client information has been input into self-learning AI tools without client consent. The buyer should warranty that they have the technological competence to supervise the inherited AI tools — a requirement under Opinion 705's competence framework.

TRAIGA-Specific Indemnification

The purchase agreement should include indemnification provisions specifically addressing TRAIGA violations — both for pre-closing conduct (the seller's liability) and for post-closing conduct during any transition period (the buyer's liability). Given TRAIGA's 60-day cure period and the AG's exclusive enforcement authority, the indemnification should account for the cost of cure efforts.

Data Migration Plan

The deal should include a detailed data migration plan covering: export of all client files from AI platforms, deletion requests to vendors, transfer of vendor accounts and DPAs, and a client communication strategy regarding AI-stored information. This plan should be a condition to closing, not an afterthought.

Regulatory Change Provision

Given the active federal preemption efforts against TRAIGA and the rapidly evolving AI regulatory landscape, the deal should include a provision addressing what happens if TRAIGA is modified, preempted, or supplemented by new state or federal law during the earn-out or transition period.

AI Tool Valuation

The practice's AI infrastructure — vendor contracts, compliance documentation, trained staff, and integrated workflows — has value that traditional valuation methods do not capture. An asset-based approach that values only tangible equipment will underprice the practice. A buyer who understands the value of inherited compliance infrastructure may be willing to pay a premium; a buyer who does not may discount the practice because of perceived AI risk. Our earlier analysis of valuation methods for selling a law practice discusses how intangible assets like client goodwill affect price — AI compliance infrastructure is a new category of intangible asset that deserves the same treatment.

Key Implications for Practice

For sellers: If you are planning to sell within the next two to three years, begin documenting your AI compliance posture now. Inventory every AI tool your practice uses, confirm whether each tool trains on inputs, obtain or update enterprise DPAs, build written supervision policies, and ensure your engagement letters disclose AI use where Opinion 705 requires it. This documentation is what a sophisticated buyer will ask for — and its absence is what a sophisticated buyer will discount for.

For buyers: Treat AI compliance due diligence as a distinct workstream, not a subset of technology review. Ask for the vendor vetting files, the AI use policies, the staff training records, and the TRAIGA compliance documentation. If the seller cannot produce them, price the cost of building them into your offer — and negotiate indemnification for pre-closing AI compliance gaps that you cannot verify.

For both parties: The intersection of TRAIGA, Rule 1.05 confidentiality, Rule 5.03 supervision, and the State Bar's Opinion 705 framework means that selling an AI-powered law practice in Texas is not a traditional business sale with an AI footnote. It is a transaction where the AI infrastructure is simultaneously the practice's most valuable operational asset and its most concentrated source of regulatory and ethical risk. Deal documents that do not address this duality will leave both parties exposed — the seller to post-closing indemnification claims, and the buyer to inherited compliance obligations they did not price and may not understand.