⬛️ mental model: 4 types of ai strategies

i was in a meeting with a cto last week (intake call for ai program) he didn’t speak much but as often with devs he asked me something impactful that kind of stuck with me. "why is everyone doing ai so differently?" and honestly, after working with maybe 20+ companies on their ai stuff these last 2 yrs (including mine), i realized there's basically four ways people are approaching this whole thing.

"why is everyone doing ai so differently?"

here's what i've seen:

the spread-it-everywhere approach this is where you just give everyone ai tools and hope for the best. saw Unilever do this actually - they've got ai running in everything from ice cream flavor prediction to supply chain optimization. pretty wild. but they had this whole central team managing it, which is the part most companies forget.

the bet-the-farm approach you know how volkswagen's going all-in on their autonomous driving platform? that's this. pick one massive ai thing and throw everything at it. i watched a fintech company burn through €15m trying to build an ai financial advisor. (spoiler: they switched strategies)

the sneak-it-in approach my favorite honestly. you just make your product better with ai but never mention it's ai. like how Spotify just knows what song you want to hear next. nobody calls their Discover Weekly "ai" but that's exactly what it is.

the show-me-the-money approach only do ai if you can prove roi in like 3 months. zalando does this.. they'll add ai sizing recommendations because it immediately drops return rates. nothing fancy, just practical stuff that works.

and here's where it gets messy...

i do keep seeing companies say they want the volkswagen moonshot thing, but then they freak out when there's no roi after 6 months. or they'll hire some fancy ai team to "transform everything" but then only let them work on tiny projects with guaranteed returns.

worst one i heard of? this company tried to do the spread-it-everywhere thing without anyone actually in charge of it. just bought chatgpt licenses for everyone and called it their "ai strategy." yeah, that went about as well as you'd expect.

the companies that actually make this work? they pick one approach and stick with it. even when it's uncomfortable.

in the end, all strategies kind of want the same thing. augmented, elevated humans that can produce better quality at scale and better quantity.

but yeah, moral of the story: pick your ai game before you start playing it. mixing strategies is basically just burning money with extra steps.

david

⬛️

👋 by the way, we run a community called genai ⚫️ circle. it’s where i learn everything i share. it’s invite-only. as a subscriber of this newsletter you can join. check it out here: www.thegenaicircle.com

join us here

other posts you might have missed