Can AI Exception Handling Be The Next Competitive Advantage?
The limits of what AI can ever do
Happy Monday and welcome to Investing in AI! Today I want to talk about the emerging dichotomy of what AI can and can’t do, through the lens of two tasks that happened to me this week.
I talk to people every day who rave about ChatGPT. “I used it 30 times a day” or “I use it for everything” they say. And I wonder a lot about what they do for work because, I don’t have that experience. Here are 3 examples of work for me that I don’t think AI can do, but, if it can and I am missing something, please let me know.
I ordered a corporate card for the COO at Neurometric and it never showed up. So I spent 20 minutes on the phone with American Express trying to understand what happened. We finally realized the building we are in requires a floor number to deliver mail, and that wasn’t on the card. It was an exception handling task I had to perform myself, as I don’t know how AI could have solved it.
Our podcast "Inference Time Tactics” was taken down from Youtube for copyright violation for the intro music. Our podcast producer claims we have a license to use it, but, the automated system at Youtube has made this difficult to resolve so, we’ve tried a bunch of workarounds to fix it. Another exception handling task that I don’t think AI could fix.
I spent a bunch of time with our lawyers working through some issues for getting stock to some advisors and employees because we had a unique complication that I won’t go deeply into here. I asked AI how to resolve it and got standard boilerplate legal solutions, not a solution to this problem. So, again, it was an exception handling situation.
All of this made me wonder about competition in AI and where exception handling comes into play. If AI automates more and more tasks, most of that capability should be available to everyone in the industry. A thesis of mine for a long time has been that AI will commoditize all sorts of things as it makes expertise, automation, and assistance more widely available. What if the differentiator becomes - how do you handle the exceptions to AI?
Could companies differentiate there? You could image a high touch fast resolution company that cost more, a low touch frustrating company that was cheap, and maybe someone somewhere in between. You can imagine, in a world where automation drives more usage, that exceptions happen more often, and handling those appropriately goes up in value.
Maybe the way to build your AI strategy is to take the best of breed AI for most things and invest in exception handling as your unique advantage. What do you think?
Thanks for reading.
Very interesting. In enterprise settings, there are a million edge cases and processes that would probably be of need for exception handling. I suspect there can be a niche market for this kind of concierge service, especially as executives are planning for AI to run across the enterprise. There may be a need for an upgraded version of a helpdesk.
Perhaps this also creates opportunities to offer more products/experiences that are purpose-built to be exceptional in a way that excludes AI agents. In essence an extension of the artisanal or organic movements to more and more domains, as some parts of the customer base are willing to pay for the extra friction that comes with human labor.