The prerequisite for steelmanning is that the argument presented to you does not "stand on its own", that it has flaws that you'd normally use to reject it - but instead, you decide to patch up the flaws and see if the stronger version works better.
It's an obviously good thing to do once you stop viewing arguments as competitive endeavor, but cooperative problem solving. If your co-worker was proposing a way to solve a programming challenge and you saw some flaws in it, you wouldn't shoot them down with "oh but this is wrong here, here, and it's also offensive". You'd say, "yeah, that could work, but it would be better if we did Y instead of X, and maybe let's not do Z because of $reasons". Applied to discussions in general, that's steelmanning.
Okay, I see your point with the competition and you‘re right (the coworking example really drove that home). Still what you describe would fall under the principle of charity for me as well.
To build on your example: You‘d patch up the flaws in your coworker‘s argument (p. o. c.) and by doing that you are then able to derive new insights on the issue (just the normal function of a discussion)
Here is the difference between charity and steelmanning. Most real world arguments are not fully specified like a philosophy book because people have limited time to fully lay out all their assumptions, to treat edge case, and explore all relevant consequences of their argument.
The principle of charity is that that whenever there are two possible interpretations of the other person's words (because of limited specificity as above), you assume the one that makes their argument stronger. Eg. I was arguing with someone about veganism and omnivorous diet, and at one point they said, "aha, because you didn't explicitly say that you are against eating brain-dead humans, you must be for it." That is them NOT giving the charitable interpretation of my words.
The principle of steelmanning goes far beyond charity. You actually play the devil's advocate and reconstruct the argument, so that flaws are removed and strengths introduced. Eg. someone is trying to convince you of the benefits of nuclear power, and they use stats from 2005. You go out there and find stats from 2018 that even more strongly support their viewpoint. So in your opposing argument, you criticize their argument as if they had quoted the 2018 stats.
It's an obviously good thing to do once you stop viewing arguments as competitive endeavor, but cooperative problem solving. If your co-worker was proposing a way to solve a programming challenge and you saw some flaws in it, you wouldn't shoot them down with "oh but this is wrong here, here, and it's also offensive". You'd say, "yeah, that could work, but it would be better if we did Y instead of X, and maybe let's not do Z because of $reasons". Applied to discussions in general, that's steelmanning.