Lexis+ AI (Lexis+ with Protégé)

Published · Researched 2026-09-22

Lexis+ AI (Lexis+ with Protégé) — AI legal research

Maker: LexisNexis (RELX). Unlike CoCounsel, LexisNexis built its AI internally rather than acquiring it. Purpose: Conversational AI research assistant layered on the Lexis content library — verified legal research with Shepard's citation validation, conversational search, summaries, drafting, and document analysis. Current 2026 version: LexisNexis replaced first-generation Lexis+ AI with Lexis+ with Protégé in February 2026, expanded it in May 2026 with a skills and orchestration layer, document vaults, and shared workrooms, and rebuilt the orchestration again in August 2026 (per Codebridge, checked 21.09.2026). Components: Protégé (conversational assistant), Create+ (Microsoft 365 drafting), Pinpoint (clause/issue identification), Shepard's At Risk (citation validation), Code Compare (statutory comparison), Search Term Maps (data visualization).

Social standing

Platform pulse:

• LinkedIn: Frank Ramos (~June 2026) picked Lexis over CoCounsel: "LexisNexis took the harder, longer road — they built Protégé internally. They understood their data, their workflows, and where AI fit. That ownership shows in the product." He also flagged the price edge and the Harvey–LexisNexis strategic alliance (June 2025): "Harvey users now pull LexisNexis primary law and Shepard's Citations directly. If your firm uses Harvey, that's a meaningful integration."
• Web community: growlaw.co's 2026 ranking gives Lexis+ AI 5/5 — "Known for authoritative, citation-verified results in high-stakes legal research." thebestcommander.tistory.com's 2025 comparison notes the interface as "intuitive, thorough, easy to navigate," with some users finding "interface complexity due to feature breadth."
• Critical counterpoint: A Canadian law professor's hands-on test (Practice Source, ~early 2025) gave Lexis+ AI "a failing grade": it returned the headnote of an entirely different case (Charkaoui) when asked to shorten a summary of Reference re Senate Reform, and when asked for the criminal-law causation test it "correctly focused on causation in criminal law… However, its response was very basic, and the cases it cited were not the leading authorities. It preferred to cite lower court decisions rather than the leading Supreme Court of Canada jurisprudence." Verdict: "I cannot recommend this to my law students, and I would not use it for my own legal research at this time." NOTE: this tested the first-generation Lexis+ AI, which LexisNexis retired in February 2026 — the finding is stale against Protégé.
• Independent accuracy evidence: Magesh et al., Journal of Empirical Legal Studies (tested May 2024, published 2025) — Lexis+ AI hallucinated on over 17% of queries, vs. ~33% for Westlaw AI-Assisted Research and 43% for GPT-4. LexisNexis was marketing hallucination-free citations at the time. Codebridge's caveat stands: "The Lexis+ AI that Magesh tested no longer exists… no independent testing covers the current product."

Recurring praise: citation-verified answers via Shepard's; internally built (not acquired) so data/workflow ownership shows; strong for firms already committed to the Lexis library. Recurring complaints: pace of change itself — the product evaluated in a pilot may differ from the one deployed (three major revisions in 2026); older accuracy evidence doesn't transfer to Protégé; interface complexity from feature breadth.

Verified quotes (platform + context):

• "LexisNexis took the harder, longer road — they built Protégé internally… That ownership shows in the product." — Frank Ramos, LinkedIn (~June 2026).
• "Given its current limitations, I cannot recommend this to my law students, and I would not use it for my own legal research at this time." — Canadian law professor, Practice Source (~early 2025), on first-gen Lexis+ AI (now retired).
• "The published hallucination figures apply to the predecessor, and we would not attribute them to Protégé." — Codebridge (21.09.2026).

Social validation: Tier 2

Two independent social/community sources with identifiable users: LinkedIn (Frank Ramos) and the named law-professor review (Practice Source). Corroborated by growlaw.co and the Magesh peer-reviewed study. Threads: no Lexis+ AI discussion found. Facebook Groups: none found (Sept 2026). Instagram: not surfaced.

Feature table (2026)

Pricing (third-party reported, UNVERIFIED by vendor)

• Estimates put Lexis+ with Protégé AI seats in a ~$250–$500 band (Codebridge, 21.09.2026); described as "generally more competitively priced than Westlaw Precision with CoCounsel" (Frank Ramos, LinkedIn).
• growlaw.co lists "~$17,500 annually" — single source, UNVERIFIED, basis unclear (per-seat vs. firm).
• Quote-only; no published pricing.

Integrations

Microsoft 365 via Create+; LexisNexis ecosystem; Harvey (LexisNexis content feeds Harvey via June 2025 alliance).

Support reputation

Enterprise vendor support via LexisNexis; no verifiable public support-rating data found — UNVERIFIED.

Community verdict

The practitioner's "built not bought" pick: Lexis+ with Protégé wins head-to-head comparisons on ownership of its data and citation verification, and it's usually the cheaper incumbent seat. The honest caveat is velocity — three rebuilds in eight months means you're buying a roadmap as much as a product, and the only peer-reviewed accuracy numbers in the market describe a product LexisNexis has already retired.