Legal technology has evolved rapidly in recent years. With hundreds of legal technologies out there and each trying to solve its own problem, how do you choose one technology over another?
The contract management space has evolved drastically in recent years. COVID has accelerated the rate at which companies and firms evaluate how they can streamline workflows, save money, and better serve their customers. How can you decide which technologies will achieve those goals? What problem are you trying to solve? What should an organization know prior to choosing a solution? What makes one legal technology product different from another? How much hands-on time is required to implement and train the tool? Answering these questions can take significant time and effort.
It may help you decide which technology to choose by first understanding two software design approaches, 1. Top-down and 2. Bottom-up. The most common approach by legal technology software providers is the top-down approach which relies on a single “industry-standard.” Top-down approaches that operate with a standard set of rules may be easier to implement in the short-term, but a bottom-up approach, which learns from how a client’s data and how they’ve done something in the past, results in a more viable, adaptable, efficient, and effective result.
Top-Down vs. Bottoms-Up: What are the differences?
Top-down approaches lack customization. It is a one-size-fits-all approach, with little, if any, customization. As a result, the tool cannot infer client-specific rules due to its reliance on generalized data rather than client data. As a result, top-down tools can address shallow problems versus bottom-up tools that address client-specific problems. For example, in contract review software, top-down tools will indicate whether the text is close to industry-standard language or will insert boilerplate text when a contract section is missing. The top-down approach fails to replicate how each client specifically treats risk and fails to abide by the client’s playbook.
While on the other hand, contract review software built with a bottom-up approach uses a client’s historical contract data to build a customized product that continuously learns and adapts to a client’s specific needs. A client’s historical data itself identifies the problem and the tool becomes more intelligent over time. The machine learning algorithm learns the rules from the data without committing them to code, which allows for customization based on the substantial historical context and an individual company’s playbook—resulting in far less hands-on training time from attorneys.
Top-Down vs. Bottoms-Up: Contract Review Process
In the contract review process, there are some scenarios where the top-down approach is useful. Examples include searching a contract for a specific term for agreement or governing clause as these provisions are extensively used and rarely changed, and are easy to create prescriptive rules around.
However, most aspects of contracts are not as cookie-cutter. Items in contracts that require nuanced editing are more challenging to create prescriptive rules. Examples include inserting company-specific sentences about compliance with a specific policy. The top-down approach does not work in cases where client-specific editing is required. A one-size-fits-all approach does not generate the work product expected from a legal department from clients.
A top-down approach is also limited to the problem it was built to solve. For example, suppose you want to add new use cases. In that case, this can require significant time and resources as it means finding a sizeable generic dataset that demonstrates the specific problem and then building new rulesets around this data.
The Clear Winner: Bottom-Up Approach
Bottom-up tools allow for more customization and improved results because it constantly learns from a client’s data, and changing the rules only require changes to the source data. This results in minimal up-front human involvement as the algorithm teaches itself. However, legal teams still need to oversee the learning process.
A bottom-up model must have access to quality data to work correctly while also requiring good gatekeepers to maintain and manage the data being analyzed. Human controls are necessary to allow users to tweak how the system learns and what it throws away. Without this, bad edits will get into the model.
The most important advantage of a bottom-up approach is the tool’s ability to evolve without human intervention. This can mean the system can grow with a company and increase ROI without a heavy lift on their part. By simply fixing a document, changing terms, or a governing law clause in a contract, the system will learn and create a new rule. These rules are customized based on your company’s dataset, which can provide results tailored to your situation, industry, or company preference.
A top-down approach has limited customization, which is essentially like repurposing technology to solve an adjacent problem. While customization is possible, it can only be within a specific set of parameters. It is impossible to create a single rule set that would work for every client situation.
If you want to invest in a contract review software with a strong return on investment, you will want to ensure that it uses a bottom-up approach. While top-down products may seem appealing, the reality is every company’s situation is unique and requires client-specific customization. The only solution that can provide this is a bottom-up approach.
