Programming

Honestly, I'd sit down with the product folk(s) and see what their thoughts are. As devs, we generally decide how moreso than what. Ask what some pain points are.

My gut feeling when I hear about an AI contractor who makes a vibe coded executive report is pretty abysmal. It feels very snake oil salesman, but I think a lot of contractors*(like the subject matter expert types you bring in, not "non FTE" contractors)* are sort of like that just out of necessity. They need to market themselves as being useful moreso than actually being useful.

But that's pretty pessimistic. I wouldn't tell your coworkers that (unless they also seem to have the same impression lol). Executive reports are useful for marketing. Does your business need to market itself? Sort of the inverse snake oil salesman maneuver lol. I'll give an example,

My company does some cyber security stuff. Along with the product that provides a way to monitor, we offer a team that monitors as well. One issue brought up at our last quarterly planning was that many clients have difficulty understanding the value of our product. An example was that one person said "well y'all aren't sending us any alerts, the last folks sent us a lot. What are you even doing?" Well, it turns out the answer is that we filter many false negatives out so they don't even end up in any sort of review queue. But they don't see that, they just see queues with little to nothing. So a goal was to get a nice executive report dashboard that showed this information in a way that's easy to understand. That way if some, perhaps aloof, executive gets the idea that "we don't get any alerts to review," they'll also have some tool that very easily shows them "of X alerts fired, Y% were deemed false positives by our team, Z remaining went into the queue."

Now, none of that is AI necessarily, but it's also something that I as a dev would never really think of. At least where I have worked. I don't work very closely with customers, and when I do they are often "internal customers" who I can be more informal with.

I'll just second what multiple others have said as a footnote, but do try to be aware of costs. Tracking AI costs should be something you try to do early. Like others have said, AI is heavily subsidized by investors right now. That's just the way out economy works. People get an idea, build it, woo investors, and then focus on growth instead of profitability. As long as they can convince investors, they don't need to be profitable. Once they can't convince more, they have to look into an IPO and/or becoming profitable. That's basically where we are now. I think I seen the phrase "tokenpocalypse" used.

Who knows, maybe two years from now we'll look at this AI craze the way we look at the NFT craze from a few years ago.

Congrats on the promotion though. Trust your instincts. Be honest. I think even being candid about your reservations with AI is wise.