Since ChatGPT caught the world’s collective attention in November of 2022, generative AI (GAI) has made great strides in the legal field. As vendors increasingly incorporate this technology into their products, many early applications in the legal industry focus on capabilities related to legal research and data extraction. As GAI matures, its ability to impact a broader swath of the legal tech stack is clear and we are beginning to see vendors in other product categories announce targeted GAI enhancements to their tool offerings.
Despite a growing acceptance of GAI in the legal field, many companies are taking a wait-and-see approach or slow-rolling development plans (possibly in the hope that it will all just go away). Although recent media coverage seems to have lost some of its early effusive enthusiasm for the technology, we believe reports of utilitarian (vs. apocalyptic) overhype are unfounded. While it is understandable that some might be wary, given past hyperbolic predictions about AI capabilities, we believe that the potential impact of practical applications of GAI should not be underestimated.
We should recognize, however, that we are in the very beginning stages of building out this technology to its fullest potential. The frequent releases of new GAI product enhancements demonstrate both a growing demand for the technology and its increasingly influential presence in the legal tech field. This is remarkable against the backdrop of ongoing questions about regulation and pending copyright lawsuits against major GAI providers such as Google and OpenAI.
Beyond legal research, workflow automation is likely to be one of the areas of legal tech most impacted by GAI. Given workflow automation’s ability to automate tasks, GAI has significant potential to make workflow automation more accessible and easier to implement. Early use cases revolve around the facilitation of virtual agents or “bot” style interactions, which can be leveraged to answer frequently asked questions regarding corporate policies and procedures and training materials.
While legal departments may provide copious amounts of digestible and accessible training materials, there is no substitute for getting an on-demand response when a question arises. The most impactful GAI models limit the universe of resources from which the virtual agent can draw information and respond to queries, providing greater confidence in the appropriateness and accuracy of responses and relevancy for the company employing it, while eliminating concerns around privacy, copyright, and so-called AI “hallucinations” (i.e., when the AI just makes stuff up).
These virtual agents can also redirect business users to the appropriate automated workflows pre-designed by the legal team, for example, to launch a non-disclosure agreement (NDA) or begin a conflict-of-interest check.
What is more compelling (and what will become even more so, as providers work on tighter integrations with other systems) is that the business requestor could begin engagement with the virtual agent via familiar and frequently used apps, such as Teams, Slack, or Salesforce. In these scenarios, the legal team is not burdened with answering rote questions. Similarly, business users would not waste time trying to find the appropriate training or automated workflow, and might not require the same change management efforts to teach them how to access and operate new tools. Legal teams could spend time more effectively by developing content and policies and collaborating directly with the business on strategic work aimed at driving revenue or managing business risk.
Our long-term vision for GAI in the workflow automation space revolves around the proposed use of “independent agents” that can perform a range of functions with minimal instruction. To create independent agents, powerful GAI systems would be prompted to manage complex tasks by interacting with and directing other AI systems, accessing information, and driving software tools on their own. This co-piloting across AI systems could allow for greater automation, efficiency, and flexibility for workflow automation. Using GAI in this way could lead to the proliferation of general legal AI assistants that would simplify and combine the use of multiple legal tech skills and tools. This could drive greater efficiency and reduce costs during workflow design and build stages, assist legal teams and the enterprise in leveraging valuable data housed in various systems to improve scheduled and ad hoc reporting and, ultimately, enhance corporate decision-making. Independent agents could facilitate the arduous behind-the-scenes work that is currently needed to create valuable knowledge resources for a company.
GAI in workflow automation could begin the process of eliminating the need for complex technical integrations between legal tools with discrete core competencies. For example, before initiating an NDA workflow, a GAI-driven virtual agent could facilitate document look up from CLM to see whether an NDA already exists with the counterparty. If not, the GAI agent could assist in the generation of a document or begin to negotiate counterparty terms (or redlines), based on previously negotiated NDAs. The GAI agent could also escalate proposed changes to Legal if it knows that particular provisions are particularly risky. Through this process, many separate business-enablement training and complex technical integration points could be eliminated. Instead, the GAI independent agent could manage the best of breed legal tools and become a core component of workflow automation deployment, creating a de facto enterprise legal management (ELM) solution that improves efficiency across the board with a much lighter technical lift. Legal would remain firmly in control of managing the risk for the company but would empower the business to move more efficiently with appropriate visibility.
We can imagine consolidation scenarios in the not-too-distant future in which workflow automation, contract lifecycle management (CLM), playbook management, matter management, and knowledge management effectively start to converge into one space. These and other fit-for-purpose tools would benefit from the ongoing developments in the use of GAI. GAI-enabled virtual agents utilizing the benefits of workflow automation tools will become the standard way that organizations access the benefits inherent in the legal tech stack and beyond. Vendors who are unable to adjust or expand their product capabilities risk falling behind. The result will be the emergence of fewer, more versatile and comprehensive tools that become market leaders over time.
The potential for foundational changes in workflow automation and process management is clear, and it seems evident to us that we are entering a new era of innovation and productivity. As we await the development of broader capabilities, we encourage both our technology partners and our clients to chase and embrace the practical use cases, demonstrate and validate the real value they represent, and be ready to quickly scale up the AI maturity curve.
Source: Above The Law
