AIOps Is More Than Just a Chatbot

It started with a Discord bot

React ChatBotify is an open-source project I maintain, and part of that meant looking after its websites and servers. When things went wrong, the usual routine was to check services, inspect logs, and at times, the fix was as simple as a restart.

Sometime in late July 2025, I started playing with a fairly simple idea - what if I could troubleshoot the infrastructure behind React ChatBotify simply by talking to a Discord bot?

The idea was simple: let the bot help me triage issues, and attempt simple recovery steps when feasible. Or even better - set up alerts and have the bot respond to them, all without having me to ask!

Then I got distracted

As with many side projects, the idea did not go much beyond an initial implementation. Soon, my attention shifted to another project I was working on with konglyyy, Vibe Engineers, where we were exploring a different set of ideas around letting AI influence runtime logic.

The AIOps experiment was left sitting on the shelf, and for the next few months, I barely thought about it.

The bigger idea

Then, in January 2026, a conversation with ning brought the idea back in a much bigger way.

We were attending a design thinking course, and as people who enjoy exploring creative ideas, we spent plenty of time bouncing thoughts off each other. In one such conversation, the AIOps bot experiment that had been shelved months earlier came up. As we talked through the idea, she suggested taking it a step further:

What if , instead of just a chatbot, we built a self-hosted full-fledged AIOps platform?

That changed the scope of the problem entirely.

What started as a small tool to help me maintain a few services suddenly became a much bigger question- what would an AIOps platform actually need to look like?

Suddenly, we were no longer thinking about a better chatbot, we were thinking about how AI could safely participate in operations. We didn't have all the answers, but the idea was challenging enough to be exciting - and exciting enough that we wanted to find out.

What followed was a lot of conversations, sketches, discarded ideas, and increasingly ambitious versions of what the platform could become.

Over the next few months, those ideas gradually took shape as AcornOps.

From an idea to AcornOps

Having an idea is one thing; building it together is something else entirely. The scope was far broader than my earlier chatbot experiment, the questions were harder, and every answer seemed to uncover another problem worth thinking about.

But that was also what made it fun - learning as we went, revisiting assumptions, and shaping the platform around what we felt an AIOps experience should be.

AcornOps has come a long way since then, and there is much more to say about the platform itself - something we'll be sharing in greater detail in a separate article.

Looking ahead

AcornOps is still evolving, and in many ways, we're still figuring out what a good AIOps experience should look like.

There is still plenty of work to do, and also plenty more that we want to share - from how AcornOps is taking shape to the decisions, experiments, and lessons that continue to come with building it.

For now, we're excited to keep exploring where this goes. Though one thing is clear: AIOps is more than just a chatbot.