Lead story
Casepoint Applies TAR Metrics and Human Approval to New Relevance and Issue-Coding Agents
Casepoint introduced Casepoint IQ, including early relevance and issue-coding agents that carry sampling, precision, recall, and post-run validation into agent-assisted review.
- What happened
- Casepoint introduced Casepoint IQ, including early relevance and issue-coding agents that carry sampling, precision, recall, and post-run validation into agent-assisted review.
- Why it matters
- The design joins newer agentic workflows to validation practices already familiar from TAR. Review teams can evaluate whether the sampling method, quality thresholds, disagreement handling, privilege controls, audit trail, and final human decisions provide evidence adequate for a matter's defensibility requirements.
- Watch
- Workflow details and performance figures came from a Casepoint media demonstration and company statements. The agents are in alpha, the demonstration used a synthetic 10,000-document set, and the reported quality thresholds and savings were not independently validated for this brief.
Sourceslawnext.comCasepoint
Briefing item
Skadden Joins Advanced Access for Relativity's Evidence-Grounded claiR System
Skadden will evaluate Relativity claiR before its planned early-2027 general release, providing feedback on conversational analysis across RelativityOne matter data.
- What happened
- Skadden will evaluate Relativity claiR before its planned early-2027 general release, providing feedback on conversational analysis across RelativityOne matter data.
- Why it matters
- Advanced-access testing at a large litigation firm can surface practical requirements for permissions, source citations, privilege boundaries, auditability, reproducibility, and performance on large evidentiary records before broader release. Teams evaluating similar systems should confirm those controls against their own data and review protocols.
- Watch
- This is an advanced-access announcement rather than a general-availability release. Capabilities, security behavior, scale, and grounding quality are vendor descriptions and were not independently tested; eDiscovery Today identifies Relativity as a sponsor.
SourcesPR NewswireeDiscovery Today
Briefing item
Bloomberg Law Previews MCP Access to Legal Data for Customer-Built AI Workflows
Bloomberg Law is previewing MCP interoperability, including a Claude integration, so legal organizations can use its proprietary data inside their own AI environments.
- What happened
- Bloomberg Law is previewing MCP interoperability, including a Claude integration, so legal organizations can use its proprietary data inside their own AI environments.
- Why it matters
- Legal and discovery teams increasingly need authoritative external content and internal matter evidence to coexist without unmanaged copying or context switching. An MCP-based delivery model raises concrete questions about entitlements, source-level citation, query logging, data retention, matter segregation, and the boundary between publisher content and client-confidential material.
- Watch
- The MCP capability is described as a preview, and the article is based substantially on an interview with a Bloomberg Industry Group executive. Availability, supported content, pricing, security controls, and production performance were not independently verified.
SourcesLegal IT Insider
Briefing item
Harvey Raises $550 Million and Acquires Agent-Reliability Company Guardrails AI
Harvey announced a $550 million financing at a $15.5 billion valuation, while reporting also identified its acquisition of AI-agent testing and reliability company Guardrails AI.
- What happened
- Harvey announced a $550 million financing at a $15.5 billion valuation, while reporting also identified its acquisition of AI-agent testing and reliability company Guardrails AI.
- Why it matters
- The financing gives a widely adopted legal-AI provider substantial capacity to develop litigation systems, while the Guardrails acquisition places agent testing and runtime risk controls closer to the product stack. Buyers should assess whether that work yields measurable gains in source fidelity, repeatability, permission enforcement, monitoring, and incident handling for evidence-sensitive tasks.
- Watch
- Funding, valuation, adoption, and investment plans are company-reported. Artificial Lawyer reported the Guardrails AI acquisition; acquisition terms and independent evidence of resulting reliability improvements were not provided.
Sourcesharvey.aiartificiallawyer.comlawnext.com
Briefing item
Major Banks Press Outside Counsel to Share AI-Driven Cost Savings
Goldman Sachs, Morgan Stanley, and Citigroup are reportedly asking outside firms to quantify AI savings and consider pricing arrangements that pass efficiencies to clients.
- What happened
- Goldman Sachs, Morgan Stanley, and Citigroup are reportedly asking outside firms to quantify AI savings and consider pricing arrangements that pass efficiencies to clients.
- Why it matters
- Discovery budgets are a major component of litigation spend, so client demands for demonstrated AI efficiency can directly affect review staffing, vendor selection, fee structures, and reporting. Legal teams will need defensible baselines for hours, throughput, quality control, exception handling, and technology charges if savings claims are to be measured without weakening review quality.
- Watch
- The item summarizes reporting by Project Counsel Media and statements attributed to bank legal leaders; the underlying commercial negotiations are not public. Savings will vary by matter, workflow, validation burden, and fee arrangement.
SourceseDiscovery Todaymyemail.constantcontact.com
Educational only. Not legal advice. Review sources and matter-specific obligations before relying on any workflow decision.