Artificial intelligence has ceased to be a novelty in the legal world - it is becoming an everyday tool for analyzing contracts and legal documents. For in-house legal teams and law firms, the key question today is not just whether to use AI, but how to do so safely, effectively and in compliance with the AI Act.
CMS describes Harvey as “ChatGPT tailored to lawyers' needs”, which allows uploading entire bundles of contracts, asking questions about their content and receiving structured answers, e.g. tables of risks or deviations from a template.
1. Lawyers build so-called playbooks in Harvey - sets of rules that automate responses to detected contractual clauses, for example flagging unlimited liability clauses.
2. The Polish tool AnyLawyer, in turn, enables automatic detection of key clauses, risks and deviations from standard documents, as well as the analysis of whole contract portfolios within a single due diligence.
3. Importantly, responses in AnyLawyer include references to specific locations in source documents, which makes it easier for lawyers to validate the results.
In practice this means a shift from manually reviewing dozens of files to working at the level of business logic - deciding which clauses are acceptable and which require negotiation.
Automating due diligence is one of the most tangible areas of benefit. AnyLawyer allows simultaneous analysis of entire document portfolios - from customer agreements to supplier contracts - for consistency of key clauses and deviations from corporate standards.
1. Thomson Reuters CoCounsel, deployed e.g. at the Greenberg Traurig firm, enables extraction of clauses from hundreds of documents in under an hour, under lawyers' supervision.
2. Such tools support a rapid transition “from material to assessment and recommendation”, shortening the time of technical analysis and focusing experts' attention on interpretation and business negotiation.
3. Legal Tech research and presentations indicate that generative AI performs well at identifying legal risks and preliminary categorization of provisions, provided the results are verified by a lawyer.
For a company this translates into a real shortening of contract review time in M&A transactions and ensuring contracts comply with group policies.
In the context of the AI Act, tools designed specifically for working with legal documents, such as LEX AI developed with the Gdańsk University of Technology, are gaining importance.
1. The platform combines generative AI with a large database of statutes and case law - over 3.4 million records - and is built on the premise that data security is a basic requirement, not an add-on.
2. In public institution pilots, LEX AI is used, among other things, to analyze and provide opinions on documents, search for current regulations and interpretations, and assess legal risks in administrative processes.
3. In the context of the AI Act and high-risk AI systems, experts emphasize the need for clear contractual regulation of which data are passed to the system, for what purposes AI is used, and how counterparties are informed about its use.
This is also the direction for business: from “general chatbots” to certifiable, domain-specific AI systems that can be defended both in an audit and in court.
For an in-house legal department or a law firm serving business, strategic implementation is crucial today:
1. Choose tools designed for law (Harvey, AnyLawyer, CoCounsel, LEX AI) instead of relying solely on general models.
2. Build your own clause and risk playbooks embedded in the company's actual compliance policies.
3. Regulate in contracts the manner of using AI and data flows, with an eye to the impending full sanctions of the AI Act.
1. Can AI fully replace a lawyer in contract analysis? No - tools such as Harvey, AnyLawyer or LEX AI are designed as support under a lawyer's supervision, who verifies results and makes business decisions.
2. Which contracts benefit most from AI analysis? The biggest benefits are seen with large portfolios of similar contracts (e.g. sales, distribution, NDAs) and in due diligence for M&A transactions.
3. Is using AI to analyze documents compliant with the AI Act? Yes, provided the system is properly classified, risk assessed, data security measures are implemented and clients and counterparties are appropriately informed about AI use.
4. How to start implementing AI in a legal department? Best to start with a pilot on a selected process (e.g. NDA review), using a domain tool and clearly defined risk playbooks, then scale the solution to other contract types.