Machine Learning Makes Invoice Review More Effective than Ever

Legal operations professionals have heard a lot about Artificial Intelligence (AI) recently. Whenever a popular new buzzword comes along in the legal management industry, it can be difficult to know whether there’s something truly beneficial behind it or if it might be overhyped. While there is almost certainly overuse of the term “AI” in some areas of the market, the technology is indeed making a genuine impact on how legal professionals work. Wolters Kluwer's ELM Solutions is leveraging one particular branch of AI – machine learning – to improve a key component of spend management: the invoice review process.

Legal bill review is critical to legal department management as it can catch a myriad of unintentional billing errors such as misinterpretation of billing guidelines or vague billing descriptions. In discussions, our clients say that although they see the value in it, invoice review can often feel like a burden. Busy attorneys would rather practice law than focus on invoices. And in-house lawyers often think of their outside counsel as trusted partners. Having conversations over disputed invoice line items can feel like it could potentially damage to good, collaborative relationships.

With clients' challenges in mind, Wolters Kluwer's ELM Solutions introduced LegalVIEW® BillAnalyzer. In devising the best way to address those needs, machine learning technology became the lynchpin to helping clients get better invoice review results than they thought possible.

Why is Machine Learning Helpful in Bill Review?

The key to machine learning’s usefulness in a particular context is the amount of data available for the AI engine to learn from. Essentially, it's strengthened by data, getting “smarter” when it has access to more information. That’s why legal invoice review is the perfect place to start applying machine learning. Powered by the largest source of legal invoice data in the world with over $100 billion in legal spend data including granular detail on each and every line item in the LegalVIEW database, BillAnalyzer learns about historical billing and invoice adjustment patterns. This makes it incredibly accurate in predicting which line items on a newly submitted invoice are most likely to need adjustment.

Using several elements of machine learning, including grammar-based predictors, natural language processing, and self-adaptive predictors, this technology is able to identify line items that have the highest likelihood of needing attention from human reviewers. The system compares submitted line items to a set of rules based on the client-specific legal service agreement and finds instances of issues such as non-billable activities, unexpected matter and task combinations, rate and activity mismatches, and even missing information. It then ranks line items by their likelihood of non-compliance with the service agreement.

An Ongoing Cycle of Improvement

Then expert analysts, highly trained and experienced paralegals and attorneys, review the invoice, focusing the majority of their attention on those line items the AI has prioritized. Because the amount of invoice data BillAnalyzer accesses is constantly increasing, the machine learning engine only gets better as time goes on. In addition, the data set is continually informed by BillAnalyzer’s expert analysts and the actions they take on invoices. It is a system in which computerized machine learning and human expertise are equally critical, each making the other even more effective over time.

This is a perfect example of how machine learning has taken a beneficial process – invoice review – and made it even more valuable. I know that there are some professionals who worry that this type of advanced technology may eventually take over their jobs. In the case of machine learning, the technology cannot replace skilled professionals, but it certainly can be used to help them work more effectively and productively.

Visit our BillAnalyzer page to learn how this innovative solution can increase billing guideline compliance, control costs, and improve outside counsel relationships for your legal department.


About The Author

Linda Hovanec

Linda Hovanec is a seasoned legal professional who has worked in the business of law for over 25 years. In her current role, she drives innovation in the core product lines of ELM Solutions, incorporating best-in-class contextual design methodologies into product development processes and using real-world data to align product roadmaps with client needs.