Legal AI Market Signals · Week of Oct. 12, 2026

Legal AI is moving deeper into the work itself.

A corporate client pulls work in-house, a litigation workflow keeps the lawyer in charge, and California writes the rules down.

Published by Apis Legal Technologies

Legal AI Market Signals, Week of Oct. 12, 2026

Legal AI is moving deeper into the actual work of legal practice. This week, three developments stood out because they show the shift from experimentation toward real change in workflow, business model, and governance.

1

In-house legal work

AT&T is bringing more legal work in-house

AT&T's in-house legal team, LegalEdge, is using AI and internal legal operations to handle work that might previously have gone to outside counsel, including research, document review, and drafting. On Oct. 8 the company announced a collaboration with OpenAI built on LegalEdge, and its general counsel has said AT&T plans to trim its roster of national outside firms over the next two years.

Why it matters
The signal is not that outside counsel is disappearing. It is that corporate legal departments are getting better at asking which work truly requires outside counsel and which work they can now handle themselves. That puts more pressure on firms to demonstrate specialized judgment, efficiency, and value.

2

Supervised AI workflows

Harvey and LexisNexis launched a supervised litigation workflow

Their new motion to dismiss workflow reviews a complaint, proposes ranked arguments for dismissal, researches each one, and drafts the motion. There is an important checkpoint built in: the attorney approves or adjusts the strategy before the system moves forward.

Why it matters
That may be a better model for professional AI than asking whether a system is autonomous. AI performs more of the workflow. The lawyer keeps authority over the consequential decision. When you evaluate any tool, ask where that approval step sits and who owns it.

3

California AI rules

California's SB 574 takes effect January 1, 2027

The new law sets requirements for attorneys' use of generative AI, covering confidentiality, verification, citation accuracy, disclosure in certain court filings, and limits on delegating the practice of law to AI.

Why it matters
The message is becoming hard to miss: more capable AI does not mean less professional responsibility. California firms have less than three months to put verification and confidentiality practices in writing, and firms elsewhere should expect similar rules to follow.

The signal underneath

Taken together, these stories point toward the same larger shift. Legal AI is moving from "What can this tool do?" to "How should this workflow actually operate?"

Who can access the information? What can AI prepare? What must a lawyer verify? Where does human approval belong? And did the workflow actually improve the way legal work is delivered?

What to do about it this week

  • Ask your biggest client which of your matters they could imagine handling in-house, and be ready to explain the judgment you add.
  • For each AI tool you use, write down the one step where a lawyer must approve the work before it moves forward.
  • If you practice in California, start a one-page AI use policy now covering confidentiality, verification, and court disclosures.

Sources

This is general information about developments in legal technology, not legal advice, and reading it does not create an attorney-client relationship.

Not sure where your firm stands?

Our free audit walks you through where the hours are going and which tasks are worth handing to software. It takes a few minutes and you get the results immediately.