Artificial Intelligence is having a moment. From ChatGPT to Bing AI, pretty much every software company is rushing to release a chatbot feature. It seems everyone is talking about AI! The power of conversational AI seems to be infinite. 

Tools like ChatGPT feel intuitive even if you can remember a world with dial-up or rotary phones. That’s what makes the current wave of AI excitement fascinating. All business executives can imagine how a chatbot — who doesn’t sleep, go on vacation or think about unionizing — could help their business.

I wouldn’t be surprised if we see even the local coffee shop use chatbots to answer questions and take reservations. Among all the hype though, there is one question no one is asking. Should you even be using AI at all?

The current wave of chatbots is technically not even AI. If you really want to impress your peers, you should refer to ChatGPT as an LLM (large language model). Looking into how LLMs work can also give you clues to how ChatGPT answers questions and why it can go so horribly wrong. In this article, I’ll keep calling it AI for simplicity’s sake.

That being said, it seems AI can solve any challenge:

Need help creating content for social media? Check.
Need new images created? Check.
Need a new website? Check.
Don’t want to argue with your spouse? Let AI do it. 

AI seems to be the panacea to everything. Just sprinkle a little of it and everything will be better.

In reality, AI — as described today — is helpful in some situations and useless in others. The difficulty is figuring out the right situations for it. Don’t believe what software vendors tell you. From their perspective, AI is just weeks away from curing cancer and solving world peace.

To help you sort through the noise, I created a handy chart looking at two simple variables in our professional lives: importance and repetition. I believe these two elements can help us figure out where AI might be the most impactful.

Important situations that are highly repetitive: These are the biggest opportunities. Creating regular content for your marketing, creating boilerplate legal contracts and so forth fall into this category. Use AI whenever you can here.

Marketing is one of the best examples. Marketing requires consistent output to be effective over the long term. Creating the relevant assets can be a major challenge especially for small teams. AI can close that gap by helping you craft text, images and even videos in a matter of minutes.

Important situations that aren’t repetitive: These have to be determined on a case-by-case basis. AI might help your team tackle problems, but it depends on the complexity of the problem and what information you feed the chatbot.

Having AI review legal contracts could save you time, but if you only do that a few times a month, it might not be worth it. Keep in mind that AI is bound to miss important elements or even make errors. Just look at the recent weird conversations people have had with the BingAI before thinking these tools are perfect.

Unimportant situations that aren’t repetitive: These should avoid AI. You’ll simply try to automate things that don’t happen often and aren’t important.

All businesses have these kinds of tasks. It’s what we often call “admin” work. Don’t worry too much about these tasks and either try to remove them or forget about optimizing them. 

Unimportant situations that are highly repetitive: These should be discarded altogether. There’s no point in automating things that shouldn’t take place anyway. 

Asking AI to create reports or presentations that no one reads is a waste of time for everybody, including the machine. Instead, you should be thinking if these tasks are even needed or if you could find better alternatives. I shudder every time I see a 50-page report that could be summarized in two pages and a 60-minute conversation.

As your team thinks about the future, AI tools should be a bullet point in the agenda. Depending on your business, AI might have a minimal impact or a major one. The goal is to find ways to drive business outcomes such as revenue, profit, talent retention and so forth. Don’t view technology adoption as the end goal. 

AI isn’t the greatest thing since sliced bread. It has fascinating use cases, but we need to keep our two feet on the ground. The hype is deafening, but I hope the chart above gives you ideas on how to better think about AI. Maybe AI will evolve and become more like the robots in Star Wars but, in the meantime, we need to manage our expectations.