Conversational AI set to reshape legal knowledge work

Legal knowledge management is set to shift towards conversational artificial intelligence interfaces and tighter human-AI collaboration in 2026, according to Lexsoft Systems chief executive Carlos García-Egocheaga.

García-Egocheaga said lawyers will rely on AI agents as their main gateway into knowledge systems, client data and business platforms. He forecast that these tools will take on multi-step work rather than deliver lists of search results.

The comments reflect growing interest in generative AI across commercial law firms. Many firms are now testing tools that sit between document repositories and practice systems.

García-Egocheaga said conversational systems such as Copilot, Claude and Gemini will become the default interface for legal professionals. He said these agents will handle context, infer intent and orchestrate access to multiple systems.

Lawyers will describe what they want in everyday language. The AI agent will break down the request into tasks and call other specialist agents in the background.

In García-Egocheaga's view, this will mark a departure from keyword search. It will also change how lawyers think about knowledge management tools.

He described a scenario in which a lawyer issues a single instruction. The AI must locate a relevant share purchase agreement with specific geographic and jurisdictional features. The same agent must then use contract management software and CRM data to generate a new version and send the document link by email.

That scenario relies on links between knowledge systems, contract platforms and client databases. It also relies on AI agents that can sequence tasks and pass results between systems.

"Knowledge search is now merely the entry point for comprehensive, automated task sequences," said García-Egocheaga, CEO, Lexsoft Systems.

Human oversight

García-Egocheaga predicted that this shift will change the structure of legal knowledge teams. He said firms will move towards a formal human-AI collaboration model.

Under this model, generative AI will handle large volumes of document classification and curation. Professional support lawyers and knowledge specialists will review and approve the outputs.

AI tools will generate more detailed metadata on precedent banks and matter files. This metadata will include high-level taxonomies and more granular labels on clauses, risks, concepts and jurisdiction.

García-Egocheaga said this depth of classification will allow AI agents to understand instructions with greater precision. It will also reduce the risk that systems return documents from the wrong legal framework.

He said human checks will remain central. PSLs and knowledge professionals will examine suggested tags, identify gaps and correct errors before information flows into live use.

In his view, this role will shift PSLs closer to the core of legal operations. Their decisions will shape the quality and reliability of the firm's AI-assisted workflows.

Connecting systems

García-Egocheaga highlighted Model Context Protocol as a key technical piece in this approach. MCP is an emerging standard that defines how AI agents connect to multiple data sources and tools.

Under this model, MCP sits between conversational interfaces and firm systems. It provides a structured way for AI agents to call contract tools, knowledge repositories, CRM platforms, practice management software and communications channels.

Firms will use MCP to assemble AI workflows that cross system boundaries. A lawyer's single request will then trigger several linked tasks inside and outside the knowledge platform.

García-Egocheaga said this will move legal workflows away from step-by-step manual work. AI agents will run more of the process from search through to document output.

He said firms that adopt MCP-based frameworks will reduce time spent on administrative steps. Lawyers will then focus on strategic and advisory work for clients.

Workflow impact

García-Egocheaga expects these changes to influence day-to-day practice. He said research, document review and drafting will rely more heavily on AI orchestration in the background.

Lawyers will still define the objectives, constraints and final outputs. AI systems will handle the intermediate steps of gathering, classifying and formatting information.

He said this will not remove the need for judgement. Instead it will move human attention towards review and client interaction.

García-Egocheaga projected that firms will treat process automation as a routine feature of legal work. He said this will sit alongside more formalised quality controls on AI-generated content.

"The evolution of conversational AI, the integration of AI-human collaboration models, and the adoption of protocols like MCP will reshape legal knowledge management and workflow execution for lawyers. These advancements will empower lawyers as well as firms to interact with information more intuitively, automate complex processes as a matter of course, all the while maintaining rigorous knowledge quality standards. Legal professionals will be able to streamline research, accelerate document analysis and creation, and execute complex multi-step tasks with efficiency. By reducing time spent on time-consuming information retrieval and process management, lawyers will be able to focus more on the human (irreplaceable) elements of legal practice," said García-Egocheaga.