[riizade]

As a very junior developer, I once worked on a chatbot for purchasing flight tickets. You know the ones, the annoying windows that pop up when you're trying to navigate a (hopefully) otherwise perfectly usable site, the ones that are intended to replace customer support but are fundamentally unable to, the ones that you'd never trust with a transaction worth $5, much less one worth hundreds or more.

Even if this project were a rousing success by the company's definition, it would have ended up a useless chatbot used by very few people but frustrating many more.

Instead, one day we're chatting about the project, and when talking about strategy for implementation, the product manager pushes back on working too early on integrating with the actual API calls to purchase tickets. I thought having a proof-of-concept for the actual functionality would be important, so if we ran into roadblocks we could ask other teams to provide us alternative APIs with enough notice at the beginning of the project rather than the end.

The product manager said that we couldn't afford to do that, because if we did, we'd lose.

I said "what? lose? what do you mean?"

It turns out, our team was one of 3 teams competing to make the same project. Leadership wanted to make 3 fully implemented, separate systems, and the one they liked best would go live, while the others would wither and die.

It also turns out, that our team was way behind because although we had some logic set up for handling edge cases in conversations (not great by any stretch), another team with no backend developers had a beautiful UI concept that handled only exact strings with no room for deviation on the user's part. This UI concept demoed well, of course, so they were the favorites.

It was made clear to me by the team's technical lead that nobody up the management chain (including our team's direct manager) had ever written a line of code before, or knew what the difference was between backend and frontend, or a tech demo that presents well versus a working product, and so wouldn't be able to differentiate a finished product that actually purchases tickets from one that runs entirely on the frontend and talks to nothing (until launch day, of course).

I was already on edge because I had been asked to test API integrations by using my personal credit card to purchase a real flight ticket and refund it because we didn't have test cards or even corporate cards we could use.

So I resigned within a week, and my manager was furious.

This story happened probably more recently than you're thinking.

That was the most useless project I've ever worked on.