Artificial intelligence and machine learning may feel mystical and hard to understand. However, as Arthur C. Clark said, “[a]ny sufficiently advanced technology is indistinguishable from magic." Machine learning may look like magic, but under the hood, it’s easy to understand.
Let’s start by understanding traditional programming. In my first computer class, the instructor did an exercise that stuck with me. It goes like this:
“Mark, please ask your neighbor Katy to stand up and walk out the door,” says the instructor. Then Mark says, “Katy, please stand up and w..”, “Wait a minute!”, the instructor says. “Let me stop you there; you see Katy is a computer and she has no idea what ‘stand up’ means.”
“OK, so Katy, to stand up is to go from sitting to upright." “Mark Katy doesn’t know what sitting is, or upright for that matter,” says the instructor.
“Wow, so what does she know?" The instructor responded saying “Pretty much zeros and ones. However, let’s assume someone came before you and taught her what a body is, arms and legs are and what a chair is."
“Well then, Katy, to stand is to straighten your legs with the rest of your body above them."
“Now you’re getting the hang of it, though she still may not be able to follow that without additional instructions,“ said the instructor.
That is traditional computer programming. It is very strict — the computer only does what it is instructed to do and can only respond with pre-programmed responses.
What is Machine Learning?
Machine learning is different in that the computer is programmed to teach itself. It digests large amounts of data and makes sense of it using several methods.
Over the years, more and more data is being collected. There was a time when that data was manageable and humans could write code and rules to manage, analyze and understand it. Now that we have reached a point where data is pouring in by the trainload every millisecond, we need new tools.
That’s where machine learning comes in. It can analyze at lightning speed, making connections and seeing trends. It can “think" about the data, at least in a sense. Much like a baby, at first it knows very little, but as it digests more and more data it learns and grows and makes more and better assumptions.
It’s really that simple.
Now don’t get me wrong, it has taken many great minds and many years to figure out how to make computers think.
Machine Learning to Automate Legal Billing
For example, BillerAssist by EffortlessLegal uses machine learning to automate legal billing.
A law firm cannot survive without proper billing, and BillerAssist makes it “effortless" to bill clients. The app’s patent-pending Assisted Billing Review feature provides clear color codes for particular legal time or expense entries that are excessive or unreasonable. Assisted Billing Review also tells you when the total amount of time accumulated in a particular matter becomes excessive or unreasonable. This feature also provides clear alerts regarding any remaining issues as to your clients’ billing rules.
All of these features ensure compliance, and reduce manual review time substantially.
In addition, Assisted Billing Review helps you standardize your bills, by letting you know when a particular text description in a billing entry is unusual. It makes your bills more professional!
BillerAssist also learns from your changes. It uses your own data for its decision-making, but also comes with a default set of data so you can get started right away.
This means our legal billing software becomes more accurate and more helpful as you use it.
For more information, please visit our website at https://EffortlessLegal.com, or stop by our booth at the ABA Techshow to learn more about how machine learning can help your firm increase revenue and profit!
This article was first published in Law Technology Today on 2/26/2019.