A partnership for the future: generative AI and in-house legal

The legal industry is in the exploratory stage of adopting generative artificial intelligence (AI) and similar technologies (see Know how article “AI and law: reshaping the legal industry”). Effectively adopting generative AI requires a nuanced understanding of key issues, including the accuracy and reliability of AI outputs, cost-effectiveness, data privacy adherence and bias mitigation. This informed and balanced approach is crucial for responsible AI adoption. In general, in-house legal teams are adopting a methodical and cautious approach by thoroughly evaluating alternatives rather than impulsively implementing generative AI solutions.

In-house legal teams wear many hats to support the company’s legal and operational needs. Broadly speaking, this includes internal management and external counsel management. In both of these areas, a number of AI use cases have emerged, offering opportunities for in-house legal teams to enhance their effectiveness and continue progressing on their AI journey.

In-house legal teams have a wide array of internal functions, which are primarily focused on offering legal advice, aiding the company’s commercial activities and strengthening legal operations. The specific needs of these teams vary based on factors such as the size of the legal department, overarching company objectives and budget constraints.

The in-house legal team should be involved in developing AI policies and procedures. With the rise of generative AI, these will need to be tweaked as best practices emerge. However, there are several key areas where generative AI is likely to be beneficial.

All companies face risks relating to both internal and external corporate communication. This risk is greater in certain industries, for example, some companies may be required to monitor communications as part of an agreement with a regulator. Where in-house legal and compliance teams need to monitor communications, they can turn to AI to accomplish this more effectively and efficiently, which can be a game-changer as time is of the essence in many of these situations. Generative AI is useful for assembling and summarising a collection of results.

In-house legal teams are playing an increasingly pivotal role in corporate information governance, that is, the process of managing business data. While these programmes currently leverage AI-based tools, technological advancements are likely to lead to the integration of more generative AI. This progression will enhance capabilities in real-time behaviour analysis and data loss prevention, reflecting a gradual shift towards more sophisticated AI applications.

Generative AI tools to aid drafting and negotiation work best at the clause level and are not yet well suited to creating complex agreements from scratch. However, in-house commercial lawyers typically have a good collection of reference agreements that can be processed by generative AI tools to make it easier to spot outliers and see what has been accepted in the past. Where generative AI shines is in improving clause-level language, for example by prompting lawyers to make clauses more buyer friendly.

Machine-learning approaches are also useful when creating agreement templates and playbooks. The rise of large language models and generative AI continues to make these tasks less difficult.

It is likely that the best results will come from building and managing playbooks and model agreements by blending technologies. Engaging with an advisory firm that specialises in AI can be instrumental in identifying the optimal combination of tools for problem-solving or process improvement. It is worth seeking out advisers offering innovative solutions that integrate drafting aids, model agreements and generative AI, as this will facilitate faster progression to executed agreements.

Contract analysis and contract lifecycle management are also areas where generative AI will supplement, rather than replace, current approaches. For example, where a company has already done a good job of capturing the relevant fields and populating a contract lifecycle management system, it does not seem that generative AI is well positioned to add meaningful value. However, where those things have not been done, generative AI, in concert with other more traditional techniques, should help to bridge the gap.

In-house legal teams participate in the preparation of a wide array of regulatory filings. Some are simple templates, while others are nuanced and complex. Preparing complex filings can be enhanced with generative AI-powered drafting aids. Similarly, when a regulatory filing requires an analysis of business operations reflected in various tools and databases, generative AI-powered analytics can assist. Another use case is for quality control and consistency checking, both internally and across filings.

The colourful term robotic process automation (RPA) evokes entertaining images of machines scurrying about tending to repeatable processes far more effectively than people. In-house legal teams have always had the opportunity to focus RPA on their workflows, but generative AI can extend the reach, for example, where a process begins with an intake communication that must be processed to trigger an automated workflow. Generative AI is well positioned to receive a plain English email message, extract the relevant information and then launch the RPA mechanism.

In recent years, innovative law firms and corporate legal departments have been harnessing the power of chatbots to address or redirect routine internal queries. This development aligns with the capabilities of rule-based expert systems, which are most effective for narrow, technical problems with complex but well-defined rules. Generative AI is taking this a step further by enabling teams to create systems that are capable of fluently asking and responding to document-related questions, with safeguards to minimise the risk of erroneous outputs. For less critical topics, this application of generative AI is expected to significantly aid small teams in efficiently managing routine company-wide inquiries.

In-house legal teams also manage external counsel: a time-consuming task where generative AI and related technologies can be useful. For example, ensuring compliance with external counsel guidelines typically involves analysing individual time entries. Machine learning models are adept at identifying specific guideline violations. The transition from traditional word-based approaches to concept-based approaches, driven by modern large language model architectures, offers enhanced analytical capabilities. It is not yet clear if the more traditional machine learning-based approaches will be replaced by or used in conjunction with concept-based large language models, including generative AI.

In-house legal teams often need to prepare and present summaries to internal stakeholders. Generative AI holds the potential to take matter-level information from external counsel and internal sources and generate uniform summaries. These tools are best used for generating drafts that need to be reviewed and refined. While the models do not typically generate final versions, they tend to accelerate the process meaningfully.

Generative AI holds the promise of making it easier to identify and deliver relevant insights from data. This use case is more aspirational, but these tools could be helpful for interpreting data and presenting the results. When it comes to managing external counsel and matter management, generative AI may make it easier to spot issues in budget management, panel relationships, billing practices and performance management.