In law firms, AI leaves the laboratory and enters the heart of the business

It is moving from being just an experimental tool in law firms to becoming a central piece of the sector’s business strategy. After two years marked by tests, pilots and adaptations, the discussion is now no longer whether the technology will be adopted, but how it can generate measurable gains in productivity, efficiency and profitability.

This is what a survey carried out by Thomson Reuters, a company specializing in legal infrastructure, shows, indicating that 45% of Brazilian legal professionals already use generative artificial intelligence tools, while another 31% intend to incorporate them in the coming months. At the same time, pressure from clients themselves for its use is growing, as 92% of corporate legal departments and 87% of tax areas say they expect firms and consultancies to use AI to improve the provision of services.

The movement suggests that Latin American law is entering a new stage of digital transformation. According to Thomson Reuters, if in 2023 artificial intelligence was seen as a promise, in 2024 it went through an intense testing phase, and now in 2026 the focus shifts to concrete results, such as cost reduction, productivity gains, new revenues and competitive differentiation.

“What we are seeing is that artificial intelligence is no longer a trend and has become routine,” says Rodrigo Hermida, vice president for Latin America at Thomson Reuters. The survey shows around 39% say they use generative AI tools several times a day, while another 31% use the technology daily.

“The question is no longer whether AI will be used but how it can transform the way we work, serve customers and generate value.” According to the executive, practically one in every two offices in the region already uses some type of artificial intelligence tool on a daily basis, and the expectation is that adoption will continue to advance rapidly as clients begin to demand not only technical quality, but also operational efficiency.

According to him, the legal market is going through a change similar to that caused by the digitalization of processes. “Nine out of ten professionals believe that AI will have a disruptive role in the profession. But the most interesting thing is that this has already started to happen. Today, practically one in two offices in Latin America already uses some artificial intelligence tool in their daily lives.”

More than “advanced Google”

According to him, the expectation is not that AI will only function as a more sophisticated search engine. “The first use was very similar to an advanced Google, aimed at document review, research and jurisprudence. Now we are entering a second phase, in which the technology becomes part of the workflow itself”, he explains.

This change also appears in the main uses identified by the research as:

• document summarization (77%);
• content review (71%);
• legal research (68%).

But the trend is for technology to advance to more complex functions. According to Hermida, more sophisticated systems can now group similar processes, identify jurisprudential patterns, distribute tasks between teams and automatically monitor procedural deadlines. “It’s no longer just a tool that answers questions. AI starts to perform tasks within the workflow effectively,” he said.

Impacts on productivity

A Thomson Reuters study also estimates that professionals who use artificial intelligence can save up to 12 hours a week, equivalent to around 240 hours a year. “When you measure the cost per hour of a lawyer or an accountant, it is difficult to ignore the economic potential of this time savings,” he adds.

Despite rapid adoption, measuring results is still at an early stage. Research shows that only 23% of Latin American organizations formally monitor AI return on investment (ROI). Another 57% do not carry out this monitoring, while 20% do not even know whether the company measures results.

Among those that track indicators, the focus remains on internal efficiency:

• cost savings (74%);
• use by employees (41%);
• employee satisfaction (35%);
• projected revenue generation (35%).

Indicators directly linked to the customer still appear less frequently, such as customer satisfaction (26%) and generation of new business (29%). For Hermida, the maturity of adoption explains this phenomenon.

Inappropriate use is also a concern

But, if on the one hand artificial intelligence has already become part of the office routine, on the other hand the advancement of technology has also led courts and judicial bodies to discuss ethical limits for its use.

Recently, for example, the Labor Court fined two lawyers from Parauapebas (PA) R$84,000 after identifying an attempt to manipulate an artificial intelligence system used by the court in a labor case.

According to the decision, during the preparation of the sentence using the Galileu system, an AI tool developed for internal use by Brazilian courts, a hidden section was found in a document presented by the lawyers. The text was inserted in white font on a white background, making it invisible for conventional reading.

According to judge Luis Carlos de Araújo Santos Júnior, the message contained instructions designed to influence the behavior of the artificial intelligence responsible for assisting in the analysis of the case. The practice is known as command injection (prompt injection) and consists of inserting hidden commands to change or direct responses produced by AI systems.

The episode reinforces a growing concern in the legal sector about how this technology can increase productivity and efficiency, but ensure robust governance, transparency and control mechanisms.

For Rodrigo Hermida, the discussion about trust has become as important as technological capacity itself. “In highly regulated professions, almost good is not enough. You need to know where the information comes from, what bases were used and what security mechanisms exist to avoid errors or manipulations.”

According to the executive, this concern explains the growing search for specialized solutions, trained on reliable legal bases and developed for environments that require confidentiality and traceability. “It’s not enough to use any artificial intelligence. A technological partner makes a difference when we’re talking about sensitive information, compliance and professional responsibility.”

Security and precision

This concern about the security of technology, which now deals with contracts, opinions, legal strategies and confidential information, led Luchesi Advogados to choose the Thomson Reuters legal platform, according to lawyer and partner at the firm, Ellen Carolina Silva.

“What differentiates it from other solutions we have been testing is the guarantee that the platform does not share our data with the external environment, offering a level of protection compatible with the nature of legal information”, he explains.

According to her, the accuracy of the answers also influenced the choice. “It has very high accuracy in reading and locating information within documents. When we learn to correctly formulate the commands, the results are quite consistent.”

Market attracts competition

The race to apply artificial intelligence in the legal sector is already moving competitors and investors. Among Thomson Reuters’ competitors is a Brazilian startup that has just become the first unicorn focused on the legal sector. Founded by lawyer Mateus Costa-Ribeiro, Enter reached a valuation of US$1.2 billion after a US$100 million round led by the Founders Fund, with participation from Sequoia Capital and Ribbit Capital.

The movement reinforces that the dispute for the digital transformation of law is just beginning. For Hermida, however, technology will not replace professionals. “AI is a support for lawyers, tax professionals and accountants to be able to do more, in less time and with better quality. Professionals who learn to work together with these tools tend to come out ahead.”