Have you ever heard of agentic artificial intelligence or AI agents?
ADP’s Naomi Lariviere explains the basics of one of the more recent and popular forms of AI innovation.
For more information, visit ADP.com/SPARK
Transcript:
Naomi Lariviere
Everybody's talking about agents right now, and it's probably a good opportunity to sit down and actually explain what an agent is. There's lots of different types of artificial intelligence, you know, and you can include, robotic processing automation or RPA. Most people are familiar with that. You can include natural language processing, machine learning, generative AI. And now you have agents.
And you know, when you hear the word agentic, generally what we're talking about is the agents and, their ability to think, plan and act and actually execute on different things that you would perhaps want it to do.
So at ADP we define them in three ways A simple agent A simple agent is an agent that will actually do a very simple task. So take a PTO request and actually action that for you or change your address. That's a simple agent. Then you have advanced agents, advanced agents. They're smarter. And what they can do is start to string together simple agents one by one.
So you can actually make an orchestrated multi-step process. So block A plus, block B plus block C, string them all together. And it ends up to be a new workflow. So that's an advanced agent. Last but not least is an autonomous agent. And autonomous agent takes all of the steps of the prior types of agents and actually automates the entire process all the way through it.
Now to ADP, we believe that I should be an enabler. Humans are always going to lead the way. They should always be in part of the process. So when we use autonomous agents, a human will always be checking the work of an agent.
