I didn't leave the job, but I had my resignation letter written over this since I would have had to maintain it:
My former boss had an absolute hard-on for "AI" and brought in this low-bid, fly-by-night "AI" software to automate all of our processes. I'm a fan of automation in general, but not this.
This "solution" was basically a glorified macro generator that would screen scrape data from our apps and key into our other apps. Not only it was built on the absolute shakiest platform imaginable, but the documentation from the vendor outright told you to setup remote desktop services in a way that was in violation of licensing in order for it to work. The stack it ran on made a Rube Goldberg machine look like sleek, fine engineering.
I repeatedly told him this was bad software, but he persisted to the point where we nearly went to production with it.
The worst part? The applications he was screen-scraping were all internally-developed. We had access to the backend, frontend, everything. Rather than writing proper processes, he threw that piece of garbage at it.
Luckily he retired before it went to production, and the new CTO shut it the fuck down.
So, I didn't quit my job over it, but I was looking and had my resignation letter written.
He made this demo he was so proud of. Watching it interactively, it was like 70 steps of "move mouse {X,Y}, click, copy, etc". I could literally hear Yakkety Sax in my head as I watched it bumble through.
After that, I went back to my office and wrote a 30 line Python script that accomplished the same thing, only sanely and with the ability to handle errors. He preferred his method since "it's easier for our non-technical folks to automate their stuff this way".
That was the exact moment I started looking for a new job.
Before I replace it with something that won't catastrophically collapse when the wind blows the wrong way, I get some sort of sick satisfaction out of doing autopsies on the house-built-of-matchsticks "solutions" that users come up with and I don't know why. Some of them are truly fascinating and make you wonder how someone could possibly arrive at that conclusion based on what they were actually try to achieve.
It's also why if I'm asked to implement something, my first question isn't "When does this need to be done?," it's "What exactly is the problem you're trying to solve?"
What a user asks for and what they actually needvery rarely intersect.
I wish I could hire you and a couple other people who replied to this lol. "Match stick architecture" is definitely something we have and I have been trying to shore up / replace for years.
Sorry, I missed this comment. I actually love doing that kind of shit, I get some sort of weird pleasure out of fixing chaotic stuff like that. That tends to be my role almost all the time; I'll come in, stay a few years, fix everything and get bored, and then move on somewhere else to do it again.
My current job is the only place that I haven't done that, because it's probably the best company that I've ever worked for.
Ah yes, my last company bought into that crap. They called it RPA for Robotic Process Automation and they also used it to access internal apps that we had full control of.
It wouldn't have been so bad if they just used it to enter data into third party websites which had no APIs or integrations.
At one point we updated the title of an HTML page and we had to revert the change because the RPA team said it would be a three week turnaround to fix their script.
I noped out of there not long after, it was yet another "project management driven" company where managers and project managers were repeatedly duped by vendors and outsourcing firms instead of hiring and retaining developers.
Oh god, I'm so sorry. That's exactly what we went through, and yep, same thing. Changing even the tiniest element in the UI would break their whole "automation".
That was one of the many things I warned about but was overridden. Lol, thankfully I was saved by his retirement and new CTO agreeing with me.