Artificial intelligence (AI) remains a subject of enduring fascination and excitement for businesses, however with a touch of intimidation mixed with curiosity, especially within the legal community. Remarkable advances in Generative AI tools like ChatGPT, DALL-E, and Synthesia have showcased their ability to produce text, images, and videos that closely resemble human creations. As a result, numerous executives are actively pursuing opportunities to harness Generative AI systems such as GPT-4 within their organizations.
The UiPath Global Legal Department, consisting of forty-four legal professionals, has consistently regarded AI as an asset and enabler of more sophisticated legal automation. In simpler terms, it serves to enhance the intelligence of software robots within the realm of Robotic Process Automation (RPA). AI-driven automation liberates individuals from mundane, repetitive tasks, allowing them to channel their efforts toward realizing their creative and strategic potential. Lawyers, in particular, can shift their focus to legal matters rather than being bogged down by routine work.
AI offers a cognitive boost to RPA robots, creating a symbiotic relationship between the two. RPA can assist in overcoming the final hurdles of AI deployment, expediting the integration of AI in production processes. According to Forrester’s projections, nearly all enterprises are expected to incorporate AI in their operations by 2025, making it one of the fastest-growing workloads on the planet. To elevate enterprise intelligence and enhance the collective IQ of the organization, scaling AI is imperative, and this is where RPA plays a pivotal role.
AI, when divorced from automation, resembles a brain without a body – incomplete. The key lies in effectively merging potent Generative AI with specialized AI models tailored for specific purposes, and then harnessing the power of automation. Integration between Generative AI and expert models, grounded in real-world business data, empowers automation robots to comprehend, reason, and generate content.
Unveiling generative AI: the legal asset you never knew you needed
It is not news that lawyers all over the world have been integrating technology in their day-to-day business, but the use of Generative AI in the legal world is still in its initial stages.
Numerous potential use cases have been contemplated but in the UiPath Legal Department we have started integrating Generative AI in two specific scenarios: AI driven chatbots and contract screening. We will delve into these use cases and explore how they provide valuable support to our legal department.
AI-driven chatbots
Legal teams often find themselves inundated with a multitude of internal queries, ranging from compliance concerns to contract interpretation. Manually addressing these questions can be time-consuming and hinder legal professionals from focusing on more important and intricate tasks. This is where Generative AI-powered chatbots step in, revolutionizing how legal teams operate.
At UiPath we introduced a legal chatbot almost five years ago. However, this technology required a group of lawyers to compile a list of essential keywords and responses. Furthermore, these responses had to be manually revised whenever a policy underwent an update. While this endeavor demanded less effort compared to individually addressing all internal queries, it still placed a significant burden on the legal department.
And this is where AI-driven chatbots jump in to help our legal teams. Generative AI chatbots can be trained using a company’s internal policies, playbooks, historical legal data, and even responses previously supplied by legal professionals within internal communication platforms or ticket systems. The process for the legal team is streamlined: they simply furnish the pre-drafted policies, established methodologies, and other informational assets at their disposal to the machine learning model responsible for educating the AI-driven conversational agent.
This streamlined training process can be performed by just one person in a matter of minutes. Consider the scenario where an internal policy is updated and instead of revising a list of fifty answers tied to potential questions and key words from the database from which your chatbot is extracting the answers, one simple action suffices. You just drag and drop the updated policy into the designated folder utilized for the machine learning model responsible for the education of the AI-driven conversational agent.
Similarly to the old chatbots, the deployment of AI chatbots does not replace the role of legal professionals, but rather empowers them. Routine tasks like answering common legal queries can be offloaded to the chatbot, thus allowing legal experts to concentrate on complex legal matters that require their specialized training and expertise while having the benefit of being easier to update.
Contract screening
Generative AI employs deep learning algorithms, neural networks, and natural language processing to decipher complex contract clauses with great accuracy. By training the AI model on contracts, legal documents, and precedents, the system learns to recognize patterns, clauses, and relevant terms with remarkable precision. So, it would be a pity not to use its capabilities for contract screening and business user self-service.
The essence of utilizing Generative AI in contract analysis lies in its exceptional ability to extract specific data from contracts. Under a normal process, if a business user wanted to know, for example, what the negotiated discounts and the expiration date in a contract are, they would address the legal team. A lawyer would read the contract to locate and record the information and provide an answer to the business owner. The entire process would normally take several hours, if not days, based on the current workload of the legal team.
Rather than solely relying on legal professionals for contract analysis, the great advantage of employing Generative AI in contract analysis lies in its ability to allow the business user to self-service for straightforward answers which do not require human and legal interpretation on cross corroboration of multiple clauses – as per the above example, the negotiated discount or the expiration date.
At the UiPath Legal Department we are entering a pilot stage to implement a feature called Ask GPT, which is currently available in public preview, and is designed to begin when a business user submits a contract to the Generative AI tool for analysis. When the business user inputs a contract, the AI model autonomously screens the document. As a next step, the business user asks to be provided with the negotiated discount and expiration date through a chatbot functionality. The AI-driven document screening tool comes back to the business owner with the answers to these two questions in a matter of minutes, also providing the text of the entire clause as evidence to back up its response.
While business users are self-serving on simple, straightforward questions, legal experts can focus on the nuanced interpretation of data within the broader legal context, applying their expertise to areas that necessitate human and legal interpretation.
Outlook
These two examples of how a Legal Department can leverage RPA combined with AI is just the very beginning of how legal practitioners will be supported by modern technology in the future. There is no excuse anymore for lawyers to delay the integration of AI in their work. We will continue to explore all technology options to make life easier, not only for the lawyers themselves, but also for all others who rely on good legal advice from a professionally organized legal team every day.