For many years there has been talk of the proverbial AI personal assistant that will act as a force multiplier in what we do, allowing us to be more productive and giving us more free time to do all of the wonderful things we want to do, such as playing computer games, texting or socializing on social media, and watching lots of videos that we won’t remember the following day.

Enter the artificial intelligence agents.  These are systems that are designed to do things for us, making decisions similar to those we might make.  AI agents can learn and adapt and once properly trained, they can perform many of the same tasks that would have previously required a human.  By virtue of their ginormous memory and phenomenal processing speed, these AI agents may actually perform certain tasks better than the human they are performing them for could.  Repetitive tasks are well suited to being performed by an AI agent—think junior employee, green but smart, who can be trained to do some of the boring repetitive work you do.  Now imagine everyone in the world had such an employee and you can see the potential of AI agents.

As technology advances and becomes every more integrated in our lives, the complexity of what we need to know to continues to increase.  AI agents may be the panacea that will give us a fighting chance at regaining control of our technology.  From the basic tasks of sifting through spam email and phone calls to calendar management to controlling our plethora of incompatible smart devices in our homes, to dealing with customer service issues for products or services we purchased or even ordering food for us, these AI agents could have a dramatic impact on our productivity and stress reduction.

Since ChatGPT burst into the public consciousness at the end of 2022, AI developments have come fast and furious.  And yet, as useful as ChatGPT, Alexa, Google Assistant, Gemini, and the ilk are, using them feels very much like you are operating manual transmission car—inefficient and a bit clunky with a lot of manual intervention required.  AI agents have to potential to turn that car into an automatic that blends performance (getting exactly what you want done) with efficiency (requiring less of your time).  As they grow more sophisticated and learn more about you and what you want done, they anticipate what you want and get it done before you even ask.

In the past couple of years we have seen the release of numerous large language model AIs such as ChatGPT and Gemini.  Ais now have the ability to respond in text or to create pictures or video.  They can also increasingly accept pictures or video as input and are able to communicate with a very human-like voice.  It is not a stretch to predict that we will see an increasing integration of all of these capabilities with the ability to take control of your computer and your phone, doing things that only you could do in the past.  As with other agents, AI agents will be working on our behalf and how they perform will reflect, for better or worse, on how we are perceived, so it will be important to make sure we are clear in our instructions to our AI agents before setting them out to do our bidding.

The year 2025 will be known for many things, and I believe one of them will be the beginning of the mainstreaming of AI agents.