On July 29, 2026, Trellis, the legal research platform that maintains the nation's largest database of state trial court records, announced the expansion of its agentic research capabilities through new Model Context Protocol connectors for both ChatGPT and Claude. The announcement, reported via PR Newswire, allows attorneys to query over 2.5 billion state trial court documents, 3,500 courts across 2,500 counties in 45 states, directly within the conversational interfaces of the two most widely adopted AI assistants. For personal injury law firms, the Trellis expansion represents a significant leap in how AI-native legal research can operate, transforming state trial court data from a separate, siloed database into an integrated intelligence layer that attorneys can access without leaving their drafting environment.
The technical architecture of the integration is what makes it distinctive. Rather than requiring attorneys to export search results from Trellis and paste them into ChatGPT or Claude for analysis, the MCP connectors allow the AI assistants to directly query the Trellis database, retrieve relevant documents, cross-reference verdicts and motions, and verify citations in real time. This means that an attorney can ask a natural-language question such as 'Find me factually similar slip-and-fall cases in California retail stores where the plaintiff recovered more than $500,000' and receive a structured memorandum complete with case summaries, judicial pattern visualizations, and verified citations, all generated within the same conversation. The agentic layer does not merely retrieve documents; it plans research strategies, searches billions of records, analyzes patterns across verdicts and motions, and synthesizes findings into actionable intelligence.
The strategic significance for PI practice is substantial. State trial court records have historically been among the most difficult and expensive data sources to access for legal research. Unlike federal court dockets, which are available through PACER, state trial court records are fragmented across thousands of county courthouses with varying levels of digitization, making comprehensive case research prohibitively expensive for most PI firms. Trellis has aggregated this data into a single searchable corpus, and the MCP integration means that PI attorneys can now leverage AI to conduct comparative case analysis, verdict pattern research, and judicial behavior modeling at a scale that was previously available only to the largest insurance defense firms and corporate litigation departments. The ability to find factually similar cases, a critical input for damages modeling and settlement negotiation, is now accessible through a conversational interface that requires no specialized training in database query syntax.
For personal injury law firm leadership, the Trellis MCP expansion carries three practical implications. First, the integration of state trial court data into AI assistants means that the most productive research workflows will be those that operate within the drafting environment rather than requiring attorneys to switch between applications, and PI firms should evaluate whether their current research tools offer similar seamless integration or whether they are creating adoption friction that limits attorney utilization. Second, the agentic capabilities of the Trellis connectors, including the ability to plan research strategies, cross-reference results, and verify citations, address the hallucination risk that has made many firms cautious about AI-assisted research, because the AI is grounded in a verified, structured database rather than relying on its training data for legal authority. Third, the availability of state trial court data through AI assistants will likely compress the information asymmetry that has historically favored insurance defense firms and large corporate defendants, which have had greater resources to access and analyze state court verdicts. PI firms that adopt these tools early will be able to enter settlement negotiations and trial preparation with more robust comparative case intelligence, improving their ability to justify damages demands and to identify the judicial forums most favorable to plaintiff outcomes. As Trellis expands its connector ecosystem to additional AI platforms, the legal research landscape is shifting from a fragmented, vendor-specific model to an open, interoperable architecture in which data and intelligence flow across the tools that attorneys already use, and PI firms that position themselves within this architecture will capture significant competitive advantages.
