AI and BAS: Hype, Help, or Both?
What AI Means for BAS Techs
AI is everywhere these days. It’s in every app I open and on nearly every website I visit. Half the companies I invest in use it as a selling point in some form.
It’s fair to say that AI is going to be a part of our day-to-day lives for the foreseeable future. But to what extent? And for the controls professional, what kind of impact will it have? Are the machines coming for our jobs?
If you’re doing install work, I imagine it’ll be a while before anything like that happens. If you’re programming, it might be closer than you think. But in short, I still think the answer is no.
Pack it in, everyone. Time to go home! Article over! Sadly, my editor insists we keep going, so let’s dig a little deeper.
Why We Aren’t Going Anywhere Soon
Don’t get me wrong, what these large language models (LLMs) can do is incredible. Having the world’s knowledge served up in response to a single question feels like the second coming of Google.
But just like Google, these tools don’t do the work for you. If you’ve ever used an AI model for something technical, you’ve probably seen it get things wrong. And if you can’t spot those mistakes, that’s on you.
For anyone unfamiliar with how AI works, here’s the simple version: models like ChatGPT are trained on massive amounts of text and use patterns to predict likely responses. Think of it like making a stew: if you have fresh, high-quality ingredients, the result will probably be delicious. Mess up the ingredients, and the stew will be more “oops” than “mmm.” The quality of what you get out depends heavily on the quality of what went in.
Now, if you’ve ever gone digging through the internet while troubleshooting, you already know the challenge. A lot of the information about our industry is outdated at best or flat-out wrong at worst. Useful resources can be hard to find, and much of the best material isn’t publicly available.
So what does that mean for AI in the BAS world? At the very least, it means you should always double-check AI’s output before putting it into practice.
My Experiences with AI So Far
In my experience, I haven’t seen much adoption of AI at the ground level in our industry. It’s not out there writing controller-level programming, configuring supervisors, or pulling wire and terminating controllers. Its influence so far has been subtle and mostly behind the scenes.
Where AI does show up is in the background, like an email about a new energy efficiency project. So far, its primary function has been informing decisions. All of that data from thousands of sensors across every room, on every piece of equipment. Sensors tracking temperature, airflow, energy consumption, and more can all be processed quickly by AI to spot inefficiencies. These insights often form the basis of decisions like launching an energy efficiency project, highlighting areas that humans might take days or weeks to analyze.
AI is far better than humans at recognizing trends in large datasets, and that’s where it’s focused. Beyond trend analysis, AI’s impact has been modest. That’s been true in my own experience and in conversations with colleagues. When it comes to actual hands-on work with BAS systems, AI still isn’t a replacement; it’s more of a supporting tool. Think of it like a helpful teammate who’s great at spreadsheets but can’t yet grab a screwdriver.
How AI Can Be Useful for Us
As I mentioned earlier, AI is best used as a tool. It’s not going to do the work for you, but it can help improve productivity in your day-to-day tasks. Think of it as a supercharged Google.
The other day, I was trying to confirm some Bacnet object IDs for a boiler integration. I’ve gotten pretty good at searching for the info I need, but using AI as a search engine, I got a link to the exact manual I was looking for, and the table I needed extracted within seconds. I didn’t have to sift through the table of contents for a handful of manuals that turned out to be for the wrong system. It sped the process up substantially.
AI is great for answering quick questions. Another way to use it is as a sounding board for ideas. When you’re troubleshooting and about to call that one coworker who always saves you in a pinch, try sending a few messages to ChatGPT first. Keep in mind, these models can be wrong, but even if they are, just explaining the problem can help you think of solutions you might not have considered. It’s like venting to that very patient coworker whose head is always in the clouds, just a different type of cloud.
Closing Thoughts
So, will AI replace BAS techs anytime soon? Not in my experience. The hands-on work we do, programming controllers, troubleshooting systems, running wires, still requires human judgment, context, and experience. AI can’t replicate that.
That said, used the right way, AI can be a surprisingly useful tool. It can save time on research, uncover trends in data, and even act as a sounding board when you’re troubleshooting tricky problems. The key is understanding its limits and always double-checking its output.
In the end, AI isn’t here to replace us. It’s here to help us work smarter. Use it to handle the small stuff, speed up repetitive tasks, and get a fresh perspective. But leave the real decision-making to the people on the ground. After all, no algorithm can replace the experience of someone who’s been in the trenches or the gut instinct that comes from it.
I’m always interested in real-world stories. If you’ve used AI in your work, send me an email, I’d love to connect.
