Legal AI foundations, discovery integrations, governed data access, and professional standards

New legal AI and discovery integrations connect research, evidence, and operational systems while preserving existing permissions.

Edition
eDiscovery Decoder Newsletter

Sent September 18, 2026

Lead story

OpenAI Introduces Astra for Law with Legal Search and Firm Controls

OpenAI introduced Astra for Law, a GPT-6 Astra configuration that combines legal-specific instructions with a search index covering U.S. primary law and is initially available through selected-firm access.

What happened
OpenAI introduced Astra for Law, a GPT-6 Astra configuration that combines legal-specific instructions with a search index covering U.S. primary law and is initially available through selected-firm access.
Why it matters
A legal-specific model configuration with primary-law search and integrations may consolidate research, drafting, and matter workflows in one interface. Legal and discovery teams evaluating it should verify source coverage, authority treatment, citation traceability, permissions, retention settings, ethical walls, and required human review before relying on outputs.
Watch
Availability is initially limited, and API access is described as coming soon. Benchmark results are company-reported from 200 questions in a private validation set; OpenAI reported 54.0% overall correctness at the highest reasoning setting, so the results do not eliminate the need to verify authorities and analysis.

Sourcesopenai.comlawnext.com

Briefing item

Everlaw Plugin Brings Permissioned Case Evidence into ChatGPT Enterprise

Everlaw released a ChatGPT Enterprise plugin that lets authorized users search and analyze evidence held in Everlaw without duplicating case files into a separate system.

What happened
Everlaw released a ChatGPT Enterprise plugin that lets authorized users search and analyze evidence held in Everlaw without duplicating case files into a separate system.
Why it matters
Direct access to structured discovery evidence can reduce file movement and let teams analyze matter records from a general AI interface. Implementation review should test permission inheritance, source-level traceability, prompt and response handling, audit logs, privilege boundaries, and whether generated work product is preserved with the matter record.
Watch
Functionality, availability, permission behavior, and traceability are described by Everlaw and were not independently benchmarked in the cited coverage. eDiscovery Today discloses that Everlaw is an educational partner and sponsor.

Sourceseverlaw.comeDiscovery Today

Briefing item

Epiq Releases MCP Connectors for Discovery and Legal Workflow Applications

Epiq announced Model Context Protocol connectors for Epiq AI, Epiq Discover, Epiq Narrate, and Epiq Access, enabling approved AI assistants to search, analyze, and act on platform data and workflows.

What happened
Epiq announced Model Context Protocol connectors for Epiq AI, Epiq Discover, Epiq Narrate, and Epiq Access, enabling approved AI assistants to search, analyze, and act on platform data and workflows.
Why it matters
MCP can make discovery data and operational actions available across multiple assistant interfaces, expanding the number of systems that can initiate searches or workflow steps. Buyers should define which tools and actions each assistant may invoke, require user-level authorization, log tool calls, constrain write operations, and test how citations and matter context are returned.
Watch
Security, governance, productivity, and no-data-movement statements are vendor claims. The announcement does not provide independent test results, detailed deployment requirements, supported-assistant matrices, or pricing. eDiscovery Today discloses that Epiq is an educational partner and sponsor.

Sourcesepiqglobal.comeDiscovery Today

Briefing item

Descrybe Adds Primary-Law Intelligence Plugin to ChatGPT

Descrybe launched a Legal Engine plugin that supplies ChatGPT with structured U.S. primary law, citation relationships, subsequent treatment, jurisdictional context, and links to source material.

What happened
Descrybe launched a Legal Engine plugin that supplies ChatGPT with structured U.S. primary law, citation relationships, subsequent treatment, jurisdictional context, and links to source material.
Why it matters
Connecting an assistant to structured authority and treatment data can improve the research layer used in litigation and investigation workflows. Teams should still confirm that cited authority is current, controlling, and factually relevant, and should test coverage, negative-treatment signals, retrieval consistency, and links to official source text.
Watch
The product description, coverage characteristics, and pricing are company-provided through an announcement carried by LawNext. The cited material provides no independent accuracy benchmark or detailed jurisdiction-by-jurisdiction coverage table.

Sourceslawnext.comdescrybe.com

Briefing item

CLOC Plans Independently Verified Legal Operations Certification

CLOC is developing a Legal Operations Certified Professional credential, with beta testing planned for its May 2027 global institute and eventual oversight by an independent board.

What happened
CLOC is developing a Legal Operations Certified Professional credential, with beta testing planned for its May 2027 global institute and eventual oversight by an independent board.
Why it matters
A common credential could give legal departments, law firms, and vendors a shared baseline for roles that oversee discovery technology, data governance, vendor management, financial controls, and operational change. Employers will need to assess how the final exam maps to practical discovery responsibilities and experience requirements.
Watch
The certification remains in early development, with beta testing planned rather than completed. The survey finding reflects CLOC's respondent pool, and the cited article does not provide response-rate, sampling, or weighting details.

SourcesLegal IT Insider

Educational only. Not legal advice. Review sources and matter-specific obligations before relying on any workflow decision.

eDiscovery Brief: Legal AI Connects to Evidence | eDiscovery Decoder Newsletter