businesswoman using her laptop while working at the office

How legal departments can advance their use of GenAI

Authors

9 minute read 09 May 2025
Related topics
Tax
Law

Four steps that can help in-house teams move from idea to implementation and realize the benefits of GenAI.

In brief

  • Many legal departments have not implemented GenAI and are missing out on potential time and cost savings.
  • Use the “read, think, write” analysis to help identify the most impactful use cases for GenAI.
  • Proactively cull and organize data to help produce better GenAI guidance.

The potential efficiency and cost-saving benefits for legal departments using generative AI (GenAI) is widely acknowledged but many have yet to fully implement the technology and are unsure how to proceed. Simply moving from the idea of GenAI to the practical application can be an elusive step, particularly when the technology itself evolves almost daily. No matter where they are on the journey, four key steps can help legal departments stay the course and confidently accelerate their ambition to harness GenAI responsibly and effectively.   

How EY can help

  • Digital law

    Our digital law team can help you identify risks and meet challenges in data and cyber, digital IP, digital regulatory law and e-commerce law. Learn how.

    Read more

1. Prepare the underlying data

All effective GenAI tools require good data, so it is critical to collect all relevant materials together in one place or a manageable number of repositories, and to take steps to manage the quality of the data. Without a well-curated data source, the best large language models (LLM) in the world will not be able to provide a reliable answer to a legal question or draft a high-quality document. Similarly, if the LLM cannot find the data, it will not be able to use it in generating output.

 

The nature of the repository itself is less important than the completeness and accuracy of the data. In general, the more data you have, and the higher quality that data, the better results you will get from GenAI tools.

 

While there may be situations in which it makes sense to be selective about which documents are kept, in most cases there will likely be value in collecting all documents of a particular type. For example, if having a complete overview of all the organization’s supplier contracts would be valuable, implementing a process which ensures that the final version of every signed agreement is stored in a single database will be important. This will allow the data to be queried and information extracted using GenAI. In the example of supplier contracts, a user would be able to identify contracts over a certain value or with a renewal date in the next six months.

 

In other instances, a more selective approach may be advisable.  For example, when collecting advice received from external law firms, large legal departments may want to introduce a selective process that filters out less important correspondence with external counsel and uploads only final advice that contains significant legal content. Legal departments can then use GenAI to query that database and identify relevant information that then only needs to be checked and updated rather than sending new, duplicate requests to external counsel.

 

When selecting documents to be used for GenAI purposes, it is critical to ensure that confidentiality requirements (such as data privacy laws) are observed and, in the case of legal documents, that privilege is maintained. This will generally not be an issue when the GenAI tool is hosted internally but will need to be addressed where a third-party supplier – or potentially even another company within your group – has access to information.

 

Maintaining the integrity of data presents its own challenges, as information – especially legal information – can go out of date very quickly. GenAI can potentially help by flagging information that may need to be reviewed or updated.

2. Become familiar with what’s possible

It’s never too soon – or too late – to start using GenAI tools. Experimentation, subject to the confidentiality considerations mentioned above, will give legal professionals a feel for what it can and can’t do.

There is no doubt that GenAI suits some tasks better than others and different GenAI tools perform differently on different tasks. Trying multiple tools and approaches will begin to give you a clearer view of the “art of the possible” that is essential to using GenAI – or any other technology or solution – in the most effective way. For example, you can ask GenAI to make a contract more favorable to the supplier or to help you analyze the impact of a recent case.

Setting up a structure to connect, support and incentivize users in this phase can be particularly effective, especially for larger legal departments. Rather than leaving users to their own devices, asking some “super users” to share experience and provide tips, or enlisting the help of the IT team with more formal training or e-learning modules will help encourage less tech-savvy team members to try out their first prompts.

Introducing an element of friendly competition or “gamification” – offering a prize for the most useful GenAI prompt, for example – may also help to encourage and incentivize engagement. For legal departments with a strong culture of innovation, linking GenAI initiatives to individuals’ personal objectives can also be an extremely effective way to accelerate progress on GenAI goals.

3. Analyze current processes and key issues

Some of the most impactful GenAI use cases for the legal department can be identified by simply reflecting on the processes that you follow, the documents that you often review or draft and the daily issues that you experience.

It may help to think of legal workflows as covering three high level tasks: read, think and write. In other words, lawyers ingest information such as a contract or a new law; they consider the implications and decide on a potential course of action; and they produce some written output such as an email advising the business what actions to take or drafting alternative wording for a particular contractual clause. Most legal workflows break down into individual components that fit into one of those categories, so it may be a useful model for analyzing where and how GenAI might support legal processes.

How EY can help

Step 1: Read

Many lawyers have become exceptionally good at reviewing large documents within a short time frame. In order to be successful, lawyers must quickly to grasp the fundamentals of a new piece of legislation, understand the implications of a contract or absorb a new business strategy. Even the best lawyers, however, cannot come close to matching the speed with which a GenAI tool is able to assimilate vast amounts of information.

 

For example, GenAI has been shown to successfully identify similarities or differences across a large group of related documents, extract key requirements meeting specific criteria from a long contract and identify potentially onerous obligations in a new piece of legislation. GenAI accomplishes all this in a matter of seconds, compared with the many hours or days that it would take manually.

 

Step 2: Think

This stage involves combining the lawyer’s collected knowledge and experience with the facts and circumstances of the present issue and considering the implications. While there are fewer use cases for GenAI – at least in its current form – in this stage, there are undoubtedly elements where it can contribute.

 

GenAI can expedite data analysis, allowing you to reach a final, more robust answer or decision more quickly. While the current generation of tools cannot draw definitive legal conclusions on their own, you can use a GenAI tool as your “sparring partner” to test for any potential biases or oversights, identify perspectives that you might have missed in your thinking or find flaws in your conclusions.

 

In another example, GenAI can ingest legislation and answer a series of questions to help analyze whether a complex legal exemption applies. A series of prompts can also be used to determine whether a contract that the business wishes to enter into would trigger a tax liability.

 

These inputs cannot replace human brainpower, but they can support it and help to reach even more informed conclusions more quickly. Furthermore, as today’s GenAI tools continue to evolve at a breakneck pace, their reasoning capabilities will improve along with their ability to support legal thought processes in a more meaningful way.

 

Step 3: Write

The production of written output based on user-defined inputs is exactly what GenAI tools are designed to do, so the potential for added value is significant here as the speed with which GenAI produces output far exceeds the capabilities of a human. Even if the output is only a first draft, GenAI still offers the potential to save a significant amount of time in the writing process.

 

For example, GenAI can summarize the key provisions of a new regulation that is relevant to an organization and the key risks that executive leadership needs to be aware of. GenAI can then use that as the basis for an executive briefing.

 

If you need to brief your business team on the options for termination of a supply contract, for example, GenAI can produce a checklist of steps that need to be taken in order to terminate the contract, with the deadlines in each case and a draft notification to send to the other party.

 

GenAI can also help to summarize key information gathered in the review of a larger volume of similar documents such as property lease agreements based on a standard form. GenAI can identify and create a draft report chronicling deviations from the standard form and thereby significantly reduce the amount of human review and drafting time required.

 

4. Test then implement

Once you have identified potential use cases following the structure and methodology outlined above, testing before full implementation to the legal department will help you iron out any rough spots. Success is likely to come from an iterative process that involves experimentation and adjustments based on feedback, discussion and refinement of ideas, and direction changes when something is not working. Some initiatives will be more successful than others, and the most impactful applications can be shared more widely within the department, along with a reminder that all output should be reviewed by a human to help ensure the quality and accuracy of the content.

Here again, some structure can help with adoption. For example, creating and circulating a list of prompts that have been found to produce good results and sharing case studies on successful uses of GenAI will not only spread the benefits of that use case, but also spark other ideas from colleagues. Quantifying the benefits – for example in terms of time and/or cost saved – will help to make the case for wider adoption even more compelling.

Move actions ahead in parallel

The actions set out above do not necessarily need to happen sequentially, and some can happen in parallel; for example, the “Prepare” step can be done alongside some or all of the others. However, the “Familiarize” step should happen before the “Analyze” step to help ensure that the team identifying GenAI use cases has a solid grasp of the potential of the technology first. 

Summary

Throughout the process, it’s imperative to remember that GenAI will not always be 100% correct. For that reason, GenAI should be positioned to the legal department as an efficiency tool to help get to your desired end result more quickly, rather than as a tool that is replacing part of what you do. Put another way, the efficiency opportunity here is not to have fewer lawyers, but rather to have better lawyers by leveraging GenAI tools effectively.

Related content

How Microsoft streamlined regulatory compliance with GenAI

This case study reveals how Microsoft legal teams are using GenAI to confidently manage regulatory change. Learn more.

How GenAI will impact the law department

In this episode of the Tax and Law in Focus podcast, the speakers explore the opportunities and challenges that GenAI will create for legal leaders.

45m 13s

    About this article

    Authors

    Related topics
    Tax
    Law

    EY refers to the global organization, and may refer to one or more, of the member firms of Ernst & Young Global Limited, each of which is a separate legal entity. Ernst & Young Global Limited, a UK company limited by guarantee, does not provide services to clients.
    OSZAR »