I recently read an article about a company that’s run almost completely with Agentic Aritificial Intelligence (AI). Agentic AI is when the AI actually performs actions and makes decisions, instead of a human. Some examples of this could include answering questions, fulfiling orders, processing claims, or handling payments. I’m sure the discovery of even more uses is coming soon.
MEDVi, specializing in personalized telehealth, was able to gross hundreds of millions of dollars in a year or two. Even more astonishing, is the company is run by one person and his brother, the only other full-time employee. There are a few dozen contractors to deal with the thousands of clients. Other than that, AI is running the show.
It’s like something from a science fiction plot to have companies run by a handful of humans and technology. Aside from the inception of the company, it’s curious to wonder what the humans do. Once the AI takes root, it could become another science fiction plot to take over the entire company. This could even include idea generation, expansions, hiring, and everything else. We’re not there yet, but it might be coming sooner than we think.
Naturally, there are arguments to both sides of using Agentic AI. On the one hand, it’s means people can “scale up” easily and quickly. I like to imagine that Agentic AI would save a lot of humans from dreary, repetitive, mundane, tedious tasks. The kind of work that technology does more efficiently and consistently. For example, auto-filing documents to the correct places. This could also include adding descriptions to the documents to make them more searchable.
However, on the flip side, it means humans aren’t necessarily acquiring skills anymore. Also, some people might not want to “agentic-AI” themselves out of a job. In my younger years, I learned a lot of valuable lessons and knowledge of how to design better systems. This is a result of the boring, tedious tasks I handled. Doing boring, repetitive tasks taught me how to create systems that would avoid all this kind of work in the future. I was also thinking about how to spare a “future” me from some of this terrible work.
Of course this was in a time before AI, agentic, generative, or otherwise. Although the shortcut is tempting, it might just be cheating us out of too much in the end.