Attendees at CES 2025 in Las Vegas. (Photo by Artur Widak/Anadolu via Getty Images)
Much of the buzz at CES and other tech conferences recently is about agentic AI and the use of AI agents to perform a variety of tasks for us poor humans. I have talked about AI agents in articles here and here.
So, it might be helpful to stop for a minute and talk about what agentic AI is and what AI agents could really do for lawyers and legal professionals. On Wednesday, a panel at CES attempted to tackle the former; I will talk some about the latter in a moment.
What Is Agentic AI? Why Does It Matter?
The presentation was entitled The Rise of AI Agents. The panel was composed of Michael Anderson, general manager of Amelis.ai, Adam Dumey, VP of World Wide Technology, Debroah Matteliano, global head of restaurants at Amazon, and Caroline Reppert, technical director, National Retail Federation.
Anderson walked us through the concept of an agentic AI, which is an AI platform that can respond to multiple inquiries in one prompt and then take action based on those prompts. It can sort thoughts and understand goals and actions. Says Anderson: An agentic agent “provides a high degree of autonomy that goes beyond traditional systems and are more intelligent, adaptable, and resourceful than other AI tools.” Agentic agents are smart, adaptable, resourceful, and can make decisions and recommendations. It’s critical that they also know when to refer the issue to a human and provide an easy way to do that.
The more important thing for most of us, though, is what agents can do for us. Anderson gave an example. An AI agent could be given the following prompt:
I need to reset my password since I want to look at my insurance policy and see how high my deductible is since I am considering a new plan for open enrollment. I am about to have a child and am curious what is the right plan to choose.
Obviously, this is a complicated question that cannot be answered by an LLM using standard Gen AI. It would require the AI to analyze the situation, look several places, and reason through the necessary steps the inquirer needed to undertake.
AI Agent Examples
Anderson suggested that an AI agent could provide the following answer:
No problem, and congratulations! I’ve sent you an email to reset your password, along with a login link to open enrollment. Your current deductible is $10,000, but as a new parent, a low deductible plan is best considering likely frequent visits. If you want to learn more about a $2,000 deductible plan, I can transfer you to our specialist.
Anderson also mentioned another thing an AI agent could do that hit home. Most of us here at CES have had trouble navigating the various venues, hotels, schedules, and the always-present casinos. An AI agent could create a schedule, make some recommendations, tell us what the timeline would work for us, and direct us where to go quickly and efficiently. All with a single prompt.
Other uses cited by the panel include ordering retail goods by sorting through options and sizes and then seamlessly ordering what we want. A personalized guide to health care along with the ability to schedule appointments and prescriptions. An AI agent that could handle customer concerns and issues: We have all screamed at the phone when some bot provides no help. But a good AI agent could perform the functions and get us to a human quickly and easily. Education is another field; having the ability to talk through what a student doesn’t understand and provide help would improve learning.
The panelists all believe that in just a few years, we will all have our very own digital personal assistant or concierge that can take care of various tasks that now occupy chunks of our time. Like ordering my coffee for me to pick up, canceling my doctor’s appointment, calling my wife and letting her know I will be late, and giving me some good take-out dinner options.
How AI Agents Could Redefine Routine Legal Tasks
But what about for law? After all, we are a human business, and people expect human interactions, right? Not so fast.
Not long ago, a study was done on responses to medical questions. A human doctor answered some of the questions, and some were answered by Gen AI. Those whose questions were answered by Gen AI were happier. Why? Gen AI was more empathetic and spoke in plain terms, addressing the real concerns of the questioner.
I can see the same thing in legal. Clients often have questions and concerns, yet lawyers talk over their heads, aren’t empathetic, and don’t take the time to address real concerns. Or if they do, the bill for the time spent is through the roof. An AI agent that could answer inquiries and then take steps like sending information, providing status, and giving real-time bill estimates. This would be valuable and would address client pain points. Client intake, bill preparations, invoicing, and routine correspondence can all be done by an AI agent.
Lawyers would also benefit from a digital personal agent. For example, a lawyer might prompt an AI agent as follows: “Agent, please open a file in the name of John Smith to prepare a will. Draft an engagement letter for me to review using standard language. Given Mr. Smith’s assets, a description of which I am uploading, please prepare a draft leaving most of the property to his children and his favorite charity. Then schedule a conference with me and Mr. Smith to go over the final documents and add it to my schedule. Oh, and send a birthday greeting to Mrs. Smith for next Monday.”
Think of the time saved. Think of the improved service. Think happier clients. It also frees up lawyers from doing tasks for which they are overqualified.
Our ethical rules, by the way, require us to bill reasonably. AI agents who presumably would not bill by the hour would reduce bills for the client and/or eliminate non-billable time for the lawyer and legal professional.
The Downside of AI Agents: Job Displacement and Inequality
Of course, as with any technology, there is a downside. AI agents will replace human administrative assistants. It will replace all sorts of work humans do. And the humans replaced will be the ones who can least afford to lose their jobs.
As a profession and society, that’s something we need to think of and prepare for. We need to look for ways to find other tasks for those displaced by helping them develop skills that could be useful in the brave new world.
Technology Has No Soul, But Humans Do
There is an electronic sign on a building overlooking the Las Vegas Civic Center that asks, “Technology has no soul. Or does it? “
Technology has no soul. Only us humans do. And we need to think about our fellow humans when we engage in technology that can do so much and displace so many.
Stephen Embry is a lawyer, speaker, blogger and writer. He publishes TechLaw Crossroads, a blog devoted to the examination of the tension between technology, the law, and the practice of law.
