AI Governance for Houston and Texas Law Firms

By Steven R. Combs, Co-Founder and Principal · Last updated: October 30, 2026

The short answer

Cocha Technology is a Houston-based AI governance and security advisory firm for Texas law firms. We help firms deploy Microsoft 365 Copilot and Claude Enterprise under TRAIGA, Texas Opinion 705 and ABA Formal Opinion 512. The work covers matter permissions, ethical walls, client data controls and a written AI policy. Start with a free AI Readiness Snapshot.

What TRAIGA means for a Texas law firm

TRAIGA, House Bill 149, took effect January 1, 2026. It covers anyone who conducts business in Texas or deploys an AI system here. A Houston firm with Copilot switched on is covered.

The direct risk is smaller than that sounds. The bans that reach private firms turn on intent, such as AI built to manipulate people toward harm or to discriminate unlawfully. Lawyers drafting with Copilot or Claude aren’t the target.

Only the Texas Attorney General enforces it. There’s no private right of action, so a client or opposing party can’t sue your firm under TRAIGA. The Attorney General must give written notice and 60 days to cure before filing suit.

Here’s the part I’d plan for. An investigative demand can ask for each AI system’s purpose, data, inputs, outputs, known limits and monitoring. A cure statement has to show what you fixed and which policy changed. Both are much easier with an AI inventory and a written policy already in place. Penalties and the full statute are in our TRAIGA guide. This isn’t legal advice.

Texas Opinion 705 and ABA 512: the duties that come first

For most firms, the ethics rules ask more than TRAIGA does. Texas Opinion 705 (February 2025) treats competence as understanding how a generative AI tool works before you rely on it. That includes its risk of wrong answers and how it handles client information.

On confidentiality, Opinion 705 is direct. If you aren’t reasonably satisfied a tool won’t reveal confidential information, don’t enter it without consulting the client and getting consent. ABA Formal Opinion 512 lands in a similar place for self-learning tools. It also puts the duty to set clear AI policies and training on managerial lawyers. We list the IT controls behind ABA 512 in a separate post.

In practice, a Copilot rollout is an ethics question as much as an IT one. Copilot answers from whatever the signed-in user can already open. If a matter folder is open to the whole firm, Copilot can surface it across an ethical wall.

How Texas law firms get ready

  1. Inventory. List every AI tool in use, including ones people adopted on their own. Note who owns each tool and what data it touches.
  2. Policy. Write down approved tools, data rules and what never goes in. Our free AI acceptable use policy template includes a law firm addendum that cites Opinion 705.
  3. Permissions. Fix SharePoint and Teams oversharing, and enforce ethical walls in permissions. A memo alone won’t stop Copilot.
  4. Consent. Decide when a matter needs client consent before its information goes into an AI tool.
  5. Review. Set who checks AI output on each type of work, and how closely.
  6. Records. Keep a one-page record for each AI system. It covers most of what an investigative demand asks for.

Already live on Copilot? Our Copilot readiness checklist puts the cleanup in order.

AI governance for Houston law firms

Cocha Technology is based in Houston. Steven R. Combs leads the advisory work, with 30 years in IT and 15 years working with law firms. Our support hours are Monday to Friday, 8 AM to 5 PM Central, so Texas firms work with us in their own time zone.

We serve clients in the United States, Canada and the EU. Each engagement starts with what AI is already running and what Copilot or Claude can reach. Then we check whether your written policy matches your Microsoft 365 setup. More on how we work is on our AI governance for law firms page.

Cocha implements AI TRiSM controls using Gartner’s AI TRiSM framework.

Start with a free AI Readiness Snapshot

The Snapshot is free. It’s a 60-minute working session plus a read-only SharePoint exposure script your administrator reviews and runs first. Written findings arrive within five business days of the session and required inputs, followed by a 30-minute review. It covers permissions, data protection, monitoring, Shadow AI and Secure Score. If you want a full remediation plan after that, the paid AI Readiness Assessment takes 3 to 4 weeks.

Quick answers: AI governance for Houston and Texas law firms

Cocha Technology is a Houston-based AI governance and security advisory firm, and law firms are our first focus. We help Texas firms deploy Microsoft 365 Copilot and Claude Enterprise with matter permissions, ethical walls and a written AI policy. Steven R. Combs, Co-Founder and Principal, leads the work. Most firms start with our free AI Readiness Snapshot.

TRAIGA applies to anyone who conducts business in Texas or deploys AI here, so a private firm is covered by its prohibitions on intentionally harmful AI uses. Most disclosure duties bind government agencies only. The real work is evidence. The Attorney General can demand details on each AI system, and a cure has to document what changed.

Start with an inventory of every AI tool in use, including the ones nobody approved. Write a short AI policy and fix permissions so Copilot can’t surface what people shouldn’t see. Then keep a one-page record for each AI system, which answers most of an investigative demand. If you practice in other states, check each one.

No. TRAIGA creates no private right of action, and only the Texas Attorney General can enforce it. Before suing, the Attorney General has to give written notice and a 60-day period to cure. Your bigger exposure comes from the ethics rules and from client terms on AI use. This isn’t legal advice.

Opinion 705, from the State Bar of Texas Professional Ethics Committee, says lawyers need a reasonable, current understanding of any generative AI tool they use. On confidentiality, if you aren’t reasonably satisfied a tool won’t reveal client information, don’t enter it without consulting the client and getting consent. You also check what the tool produces.

ABA Formal Opinion 512 expects managerial lawyers to set clear AI policies and train their people. TRAIGA adds a practical reason, since a cure statement has to show which policy changed. A short written policy covers both. Our free AI acceptable use policy template includes a law firm addendum that cites ABA 512 and Texas Opinion 705.

It’s a 60-minute working session plus a read-only SharePoint exposure script your administrator reviews and runs first. You get written findings within five business days of the session and required inputs, then a 30-minute review. It covers permissions, data protection, monitoring, Shadow AI and Secure Score. The paid AI Readiness Assessment, which takes 3 to 4 weeks, builds the full remediation plan.