I’m going to try really hard not to make this space completely about AI but it is the AI conversation on the internet that inspired me to talk about problem-solving.
Recently I saw a post about a company that used to try to sell us on robotaxis and how they’ve figured out the AI thing. They went on to explain that they deployed talented engineers at real business problems where workflows needed to be transformed — and then those talented engineers fixed the problem.
And AI was in the room with them so the story is “AI saved the day!”
I have to chuckle a bit because it’s fairly obvious to me that when you have smart people who are given the mandate to solve a problem for others (who are also reasonably intelligent and have a solid problem statement), they will crack the nut regardless of what technology they have to use — because I lived this myself.
For some background context, early on in my career I spent a couple of years at a nonprofit and then a couple of years at a large media and entertainment company. After 4 years of two different flavors of work dysfunction, I fell for the siren song of a start-up. And, one year of chaos later, I ended up back at another large media company. As it turns out, start-ups are not all they are cracked up to be for folks like myself who thrive in some degree of certainty and structure.
The next role was something of an adventure in that midway through my tenure one of my colleagues told me, “My boss has an opening and is looking for someone like you. You should apply.” I had been volun-told into new projects at prior gigs but this was the first time I was navigating a real transfer into a different department internally.
I jumped at the opportunity — in part because of what I was leaving behind (hello, misogyny in tech!) but also because of what I knew I could fix.
The ask was to come in and transform the user interface of an internal suite of products, some of which had also been licensed to a competitor (and thus represented a revenue stream for the company). While the product functionally did what it needed to do, the licensee was dissatisfied with the experience overall — and I completely agreed as I had to use these tools, too.
When I started, I was given control to start with one of our products the licensee didn’t use (but still needed improvement) as a bit of a trial of what I could do. I wireframed my approach to get approval but what they really wanted was a functional prototype — so I went ahead and built to the specs I set out knowing deeply what the pain points were.
And being deep in the problem space meant I had a lot of decent ideas which were well-received by my leadership team (and our end-users)!
Over time I earned the trust of my management to work directly with the licensee contacts. As part of a very lean engineering team where I wore many hats, I made progress on greatly improving the end-user experience.
And I even got to the point where I convinced all the gentlemen on my team that we could build a working prototype of our subscription product that, as the customer desired, was able to accept payment in the form of bitcoin — back in 2013.
That feature never launched (would’ve been a legal and logistical nightmare, I’m sure), but I know the team enjoyed the challenge and the realization of “wait a sec, did we actually pull off this crazy request?”
And to be clear, no AI was involved. Just a group of smart folks listening to and collaborating closely with their client.
AI is a tool that can help solve a problem but it is no substitute for having the right people in the right environment empowered to go after the right problem. In fact, I’d argue that going too far in prescribing what tools need to be used tend to result in solutions that are searching for a problem.