Agentic AI has come of age. Carlos Garcia-Egochiaga discusses how law firms are deploying it to maximum effect.

This year, the knowledge management function is moving towards a more advanced and refined utilisation of AI. As 2026 progresses, this technology will play a significantly transformational role.

Knowing, not searching

Conversational AI agents (Copilot, Claude, Gemini and others) are fast becoming the primary interfaces that lawyers are using for interrogating and searching knowledge management systems, information repositories and business platforms. Lawyers expect these tools to deliver complete answers – they want to “know”, not merely “search” for information.

Lawyers want these conversational AI interfaces to understand context, anticipate needs, and autonomously coordinate information retrieval, delivering complete answers or finished work, as opposed to simply surfacing search results.

This is a major departure from conventional knowledge search. Increasingly, as the technology gets more sophisticated, lawyers will give complex, multi-step queries in natural human language to AI agents which will then be expected to seamlessly execute the task – often working with multiple AI agents in the background – gathering information from across business systems and platforms to deliver comprehensive responses.

Multi-step workflow

Here’s an illustration. A lawyer issues a single natural language command: "Find me a Share Purchase Agreement that has to do with Germany, and it has England as the jurisdiction. Using the Contract Management Software, create a new version of the agreement. For this version, use the data from the CRM system addressed to Tim Magnus. When done, send a link to the contract via email to my assistant.”

Searching for knowledge now serves only as the starting point for fully automated, end-to-end task execution.

This scenario shows a major shift: the lawyer no longer needs to just retrieve the Share Purchase Agreement. They can delegate the entire multi-step workflow to the AI agent, who then must seamlessly navigate the contract management system, integrate the CRM data, generate the customised document, and deliver the results via email to the lawyer’s assistant – all from one complex instruction.

Searching for knowledge now serves only as the starting point for fully automated, end-to-end task execution.

AI and the human – smarter together

The landscape of legal knowledge management is on the verge of an overhaul, driven by the synergy between AI and human expertise. Generative AI handles the extraction, classification and curation of documents, while professional support lawyers (PSLs) and knowledge management experts supervise and review the process.

This essential human-in-the-loop model leads to a robust partnership, where AI delivers powerful, scalable processing to manage large volumes of information, and the PSLs act as quality stewards, ensuring that AI-generated outputs are refined into reliable, enterprise-grade knowledge assets.

The landscape of legal knowledge management is on the verge of an overhaul, driven by the synergy between AI and human expertise. 

This will result in AI systems generating granular metadata, extending beyond the all-important taxonomies to capture key points, risk indicators, specific clause variations, and nuanced legal concepts from across firms’ document repositories. The result? Classification at a scale and depth that has not been possible thus far, in turn enabling the AI agents to accurately understand and deliver the tasks assigned to them.

Continuing the scenario above, if the documents extracted from the knowledge management system were from the "England and Wales" jurisdiction, the lawyer could then ask the AI agent to only retrieve England jurisdiction documents. Achieving this degree of precision depends on thorough document classification, where AI handles the extraction process, but a human expert ultimately verifies that the selected documents are indeed correct.

MCP and agentic AI in action

Model Context Protocol (MCP) is emerging as the critical glue that connects multiple AI agents to perform highly defined legal tasks, spanning everything from information retrieval to task completion.

MCP acts as the default, standard link, enabling conversational AI tools to effortlessly connect with a wide range of legal and business systems within the firm, including contract management software, knowledge databases, practice management platforms, communication applications, and CRM tools.

MCP is poised to transform legal workflows from sequential, manual processes to a coordinated and automated task completion framework, fully managed by AI.

Therefore, MCP is poised to transform legal workflows from sequential, manual processes to a coordinated and automated task completion framework, fully managed by AI.

Lawyers will verbalise complicated, multi-step instructions using everyday language, and AI agents will independently carry out the entire workflow. This includes launching searches across knowledge platforms, retrieving relevant legal precedents, integrating information from various business systems, and composing customised documents – in turn delivering the final product in lawyers’ preferred format.

Fundamentally changed landscape

By implementing agentic frameworks powered by MCP, organisations will not only position themselves ahead of competitors but also enable their lawyers to minimise time spent on routine legal tasks and devote more attention to strategic matters and client service.

As conversational AI continues to advance, alongside the emergence of effective AI-human collaboration models and the widespread implementation of protocols such as MCP, the landscape of legal knowledge management and workflow execution is set to fundamentally change. These innovations will allow both individual lawyers and entire firms to engage with information in a more natural and efficient manner.

Legal practitioners will benefit from simplified research, and sped-up document analysis, creation, and review. Ultimately, by cutting down on the time devoted to tedious information gathering and administrative procedures, lawyers will dedicate more focus to the uniquely human aspects of legal work – that cannot be replaced by technology.

Visit

Lexsoft Systems

Connect with Carlos García-Egocheaga via LinkedIn