---
title: "Legal AI Vendor Governance | EFROS"
description: "Legal AI vendors scored against US AI governance frameworks — sector-weighted composite, source-cited per cell, ungated and updated quarterly."
canonical: https://efros.com/research/us-ai-vendor-governance-index/by-sector/legal/
---

- [Home](https://efros.com/)
- /[Research](https://efros.com/research/)
- /[US AI Vendor Governance Index](https://efros.com/research/us-ai-vendor-governance-index/)
- /Legal AI

By Sector / Legal

# Legal AI Vendor Governance

Generative AI tools deployed in law firms scored against ABA Formal Opinion 512 (July 2024), Model Rules 1.1 / 1.6 / 5.1-5.3 / 8.4(c), and state bar AI opinions across NY, CA, FL, DC, TX, IL, NJ.

Edition: 2026-Q2 Vendors: 6 Sector-weighted composite

By [Stefan Efros](https://www.linkedin.com/in/stefanefros-cyberdefense/), CEO & Founder, EFROS

Updated · July 25, 2026

## Why this sector view

Legal AI carries privilege risk that productivity or healthcare AI doesn't — a single prompt to a vendor that trains on inputs can arguably waive attorney-client privilege for the matter. The composite weights ABA Op 512 readiness at 2× baseline and BAA-equivalent at 1.5×.

## Primary frameworks anchored
- →ABA Formal Opinion 512 (July 2024)
- →ABA Model Rules 1.1, 1.6, 5.1, 5.3, 8.4(c)
- →NY State Bar Op #1224 (2024)
- →California Bar Practical Guidance (2024)
- →Florida Bar Op 24-1, DC Bar Op 388
- →Fed R Evid 502 / work product doctrine

| # | Vendor | Score | Grade | BAA | Opt-out | US Res | SOC 2 | ISO 42001 | NIST AI | CO AI | §1557 | SR 11-7 | ABA 512 | Subproc | TC |
| --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- | --- |
| 1 | Thomson Reuters CoCounsel | 80 | B | Yes | Yes | Yes | Yes | No | Partial | Partial | N/A | N/A | Yes | Yes | 4/5 |
| 2 | Lexis+ AI | 76 | B | Yes | Yes | Yes | Yes | No | Partial | No | N/A | N/A | Yes | Yes | 4/5 |
| 3 | Westlaw Precision AI | 76 | B | Yes | Yes | Yes | Yes | No | Partial | No | N/A | N/A | Yes | Yes | 4/5 |
| 4 | Harvey | 74 | B | Yes | Yes | Yes | Yes | No | Partial | Partial | N/A | N/A | Yes | Partial | 3/5 |
| 5 | Ironclad AI | 63 | C | Yes | Yes | Yes | Yes | No | No | No | N/A | N/A | Partial | Yes | 4/5 |
| 6 | Spellbook | 45 | D | Yes | Yes | Partial | Partial | No | No | No | N/A | N/A | Partial | Partial | 2/5 |

## Buyer's guide for this sector

For a law firm adopting generative AI, the highest-leverage scoring axes are BAA-equivalent (mandatory for matter-confidential content), ABA Op 512 readiness (the controlling ethics framework), training-data opt-out (privilege-preservation gate), and US data residency. Vendors at C or below carry significant disciplinary defensibility risk if a state bar audits firm AI use.

Operationalize the scoring

### AI Governance for Law Firms

The Index tells you which vendors clear the bar. The companion resource tells you how to turn that selection into a deployable governance program with documented evidence.

[AI Governance for Law Firms →](https://efros.com/resources/ai-governance-for-law-firms/)

## Compare Legal AI with other sector and state views

Other sector views

[Healthcare AI](https://efros.com/research/us-ai-vendor-governance-index/by-sector/healthcare/) [Foundation Models](https://efros.com/research/us-ai-vendor-governance-index/by-sector/foundation/) [Enterprise Productivity AI](https://efros.com/research/us-ai-vendor-governance-index/by-sector/productivity/) [Banking & Financial Services AI](https://efros.com/research/us-ai-vendor-governance-index/by-sector/banking/)

State views

[california →](https://efros.com/research/us-ai-vendor-governance-index/by-state/california/) [colorado →](https://efros.com/research/us-ai-vendor-governance-index/by-state/colorado/) [new york →](https://efros.com/research/us-ai-vendor-governance-index/by-state/new-york/) [illinois →](https://efros.com/research/us-ai-vendor-governance-index/by-state/illinois/) [texas →](https://efros.com/research/us-ai-vendor-governance-index/by-state/texas/)

Scoring as of 2026-05-13 from public information (vendor trust portals, BAAs, SOC report cover pages, model cards, vendor documentation). Posture changes frequently — re-verify with the vendor's trust center before contract. Methodology: [read the full methodology](https://efros.com/research/us-ai-vendor-governance-index/methodology/).

## Turn the scoring into a deployable program

The Index tells you the posture. These engagements turn the posture into operational evidence.

[Reserve $5K AI Governance Audit](https://efros.com/contact/?source=ai-vendor-index-legal) [Run AI Risk Score →](https://efros.com/tools/ai-risk-score/?source=ai-vendor-index-legal) [See full Index →](https://efros.com/research/us-ai-vendor-governance-index/)