DeepJudge’s CTO on Connecting to Claude Cowork

2026-02-09 · Source: Artificial Lawyer · Field: Legal & Regulatory — Legal Technology (LegalTech) · Depth: Intermediate, short

Summary

DeepJudge, a legal search company, has integrated its platform with Anthropic's Claude Cowork via an MCP (Model Communication Protocol) extension. This integration allows Claude to call DeepJudge for permission-aware search and synthesis across a law firm's proprietary prior matters and work product. The results are then returned to Claude for further reasoning and workflow steps, including the deployment of Claude's legal plug-ins for tasks like contract review. DeepJudge CTO Yannic Kilcher emphasizes that while LLM wrappers are becoming common, true differentiation in legal AI comes from anchoring tools in hard-to-replicate assets like proprietary legal data and a firm's institutional knowledge, context, permissions, and provenance. Kilcher notes a trend where new, more powerful common models and applications like Claude are "swallowing" older, specialized LLM-inference-centric software, making open protocols like MCP crucial for interconnected AI systems.

Key takeaway

For AI Architects and Directors of AI/ML evaluating legal tech integrations, recognize that generic LLM wrappers offer diminishing returns. Prioritize solutions that integrate specialized, data-anchored tools like DeepJudge via open protocols such as MCP. This approach allows your teams to leverage powerful common models like Claude while preserving and utilizing your firm's unique institutional knowledge and proprietary data, ensuring long-term value and avoiding rapid obsolescence of your AI investments.

Key insights

Legal AI differentiation stems from proprietary data and institutional knowledge, not just LLM wrappers.

Principles

• New LLMs supersede specialized fine-tuned models.
• Generic LLM wrappers are becoming obsolete.
• Open protocols foster interconnected AI systems.

Method

Claude uses MCP to call DeepJudge, which performs permission-aware search on firm data, returning results to Claude for reasoning and workflow integration.

In practice

• Integrate specialized tools via open protocols like MCP.
• Focus on proprietary data for AI differentiation.
• Leverage common apps for user experience.

Topics

• DeepJudge
• Claude Cowork
• Legal AI
• LLM Integration
• MCP Protocol

Best for: VP of Engineering/Data, Director of AI/ML, AI Architect, Legal Professional, AI Product Manager, CTO

Related on AIssential

• Panic Rises in Legal Industry Due to Anthropic’s AI Plugins — General-purpose AI models with plugins can rapidly disrupt domain-specific software markets and professional roles.
• Connecting LLMs to the Real World: Tool Use, Function Calling, and MCP — LLMs use an application layer and standardized protocols to interact with external tools, extending their capabilities beyond text generation.
• Lawyer Uses Claude Skills, Legal World Loses It… — Customizable LLM workflows can encode individual expertise to automate complex professional tasks.
• Don’t Choose Between Claude and Antigravity: Use This Hybrid CLI — A hybrid CLI routes complex logic to one LLM and heavy scaffolding to another, drastically cutting token costs.
• AI-Chat with Strava — Developing an LLM-Integration with MCP — Integrating LLMs with third-party APIs enables powerful, personalized data analysis and actionable recommendations.
• From Local LLM to Tool-Using Agent — Local LLMs can become powerful tool-using agents by integrating with agent runtimes and external services.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Artificial Lawyer.