If you’ve ever tried using an AI chatbot to solve a tech issue, you may have already met your first Level 0 IT Technician. If you haven’t heard about Level 0 Techs yet, it’s a term that’s being used to describe those virtual AI helpers who will gladly give you advice on any tech problem you may have. The only problem? They are often wrong.
AI tools are popping up everywhere, from your internet provider’s website to the help button inside your accounting software. They’ll greet you warmly, and they can sometimes fix a small problem. Quite often though, they will misdiagnose your issue, suggest rebooting something irrelevant, and then confidently recommend a fix that hasn’t been true for decades. It might even feel like they’re just making stuff up. Because sometimes they are.
What’s Actually Useful About AI in IT Support?
These new AI bots do have legitimate and growing potential in the IT world, especially when they are used to help with brainstorming and (perhaps) routine problems. Here are a few places where AI can pull its weight:
- Password Resets and Login Issues: AI bots can often walk users through common authentication issues quickly, especially when connected to identity management systems.
- Basic Troubleshooting Steps: Restarting services, checking internet connectivity, or walking through setup instructions can be streamlined by a well-trained bot.
- 24/7 Triage: An AI agent can capture the details of a support request after hours, and if they cannot fix it on the spot, your human technician has the information they need to dive in when they finish that annoying sleeping thing that they do.
- Documentation Lookup: Some AI tools can scan your internal knowledge base or support wiki and return helpful results — assuming the info is accurate and up to date.
If you’re running a small professional services business, adding a “Level 0” AI layer to your tech team can be tempting, but we have not turned on this feature quite yet for the clients that we serve. Why not? Our feeling is that while it can be a helpful tool for IT professionals, AI techs can easily end up wasting a bunch of time if you don’t have the background to quickly spot when they start making stuff up.
Where AI Can Go Off the Rails
Before you put your help desk on autopilot, let’s talk about AI hallucinations, which is a polite way of saying the AI sometimes just, well, lies. Not on purpose, of course. But when it doesn’t know the answer, it will confidently invent one that sounds right, even if it’s dangerously wrong.
What can happen if you let an AI agent run the show? Here are a few examples:
- AI agents have instructed users to fix printer problems by performing a factory reset. While this may (or may not) fix the issue, a factory reset is rarely necessary to fix a printer.
- AI agents have also recommended a complete reinstall of Windows to fix an error with a piece of software. We have had to take this step with a client’s computer in the past, but it is super rare, and there are usually better fixes available that take much less time.
- We have also seen instances where the AI agent recommends temporarily disabling the system’s antivirus or firewall to fix a problem. Disabling antivirus or the firewall is something that should be done with extreme caution and careful planning. Plus, if you are in a regulated industry, disabling either one may violate compliance requirements.
I could go on, but you get the idea. The key point here is that a trained IT professional would have quickly identified each of the scenarios above as a non-optimal solution and would have explored simpler, safer fixes first before taking such drastic steps.
The Right Way To Use AI in Your IT Plan
For now, I would treat an AI tech application much the same way as I would a stranger who says they are good at tech. It’s okay to ask it questions, and if you understand the answer and it’s not crazy hard to try the fix, or it’s something you can easily undo then, sure, give it a try.
Just like a junior tech who needs some oversight, using AI to gather information or clarify basic tasks is fine, but always double-check critical fixes with a real human. AI can be helpful at handling repetitive questions and simple workflows, but when it comes to cybersecurity, compliance, or high-stakes system changes, you still want a qualified human to double-check the solution before you hit “go.”
Need help figuring out when it’s safe to rely on AI and when it’s time to call a human? Feel free to schedule a free consultation, and we’ll be happy to talk through it with you.