Utilizing data to increase legal industry productivity
How the legal technology revolution is a game changer in understanding data
The legal industry has begun to adapt to changes in technology, which can streamline processes, automate repetitive tasks (such as billing) and enhance overall productivity. Artificial intelligence (AI), in particular, has become a game changer in helping firms become more efficient.
AI enables users to type in a prompt and instantly receive information collated from a variety of specified sources. The key to this process is those sources. Endless amounts of content exists, but accessing good content is not always the easiest. Some content can be from a firm’s own systems, some is from the web and some the AI might make up as it gets creative. The key to using AI correctly is to understand how the tool uses information and exactly where that data is coming from.
Artificial intelligence brings the power of automation, data analysis, and decision-making to legal processes. AI algorithms can process vast amounts of information quickly and accurately, augmenting the capabilities of legal professionals. ChatGPT, an advanced generative AI language model, is an example of how the technology can be leveraged to provide intuitive and reliable assistance to lawyers. At Helm360, we have incorporated GPT into our Termi chatbot (more on that later). Before leveraging AI in a law firm, answer this: How do we know the information we receive is truthful?
Understanding AI infrastructure
Not all data is good data, and not all data is equal. Good data is accurate and complete. To organize data from many sources into one single location and achieve a single source of truth, AI infrastructure relies on both structured and unstructured data.
Structured data is data that exists in a firm’s relational databases, billing systems, matter management solutions, spreadsheets and other internal systems. Structured data is often organized by rows and columns and is easy to search, analyze and retrieve information. Think of it like a list of names, addresses, etc.
When combined with AI, structured data can be leveraged to reveal underlying relationships between different sets of information. This approach enables firms to understand financial trends, case histories, client relationship dynamics and even make predictions to inform business decisions.
Unstructured data does not have a predetermined organization or format, and does not fit into traditional rows or columns like structured data. Unstructured data could be evidence in a case, such as text documents, emails, social media posts, audio recordings, images, videos and downloaded web pages. This presents unique challenges for analysis and interpretation because it does not have a consistent structure that can be easily processed by traditional algorithms. Unstructured data is much harder for AI to process and can sometimes even lead to fictional results from AI when it’s called upon to make guesses. Unstructured data, though lacking a predefined structure, can be analyzed using AI techniques such as natural language processing and computer vision. These methods help extract valuable information on which AI can provide insights.
Creating a single source of truth
Establishing a single source of truth is crucial when using AI. It ensures that accurate and up-to-date information is readily available, minimizing errors and enhancing decision-making. When lawyers are searching for answers, they want correct information the first time (who doesn’t?). It’s inefficient to have to fact check and double-check when someone could have correct information the first time.
When attorneys use AI to find information quickly, data is being pulled from several different places – including structured and unstructured data. For example, AI might pull accurate information from internal firm systems (such as case management software), approved outside sources such as bar association websites, or legal journals. There are also many, many inaccurate sources it can extract information from, such as unverified blog posts, untrustworthy media or random social media posts made from an average person. Without verified data to answer the query, AI can even “get creative” and make up information and return it to an attorney, such as when an attorney recently received information from a client that was completely made up by ChatGPT.
To solve that problem, Helm360 incorporated GPT into Termi in a way that clearly shows the sources it has used to arrive at results. When given a prompt by a lawyer or staff member, Termi will first search internal sources such as accounting systems, client files, emails, etc… If it finds what it needs it will provide a response based on that data. Should the search turn up empty, Termi will move on to approved external sources. Finally, if all else fails, Termi may then create an answer based on unapproved sources from the rest of the internet, letting users know.
Crucially, when it delivers results to a user, Termi will clearly show how it sourced its information. It will also indicate the degree to which it has gotten creative with the response.