AI is really good at helping solve consumer problems that weren’t being solved before. Problems caused by friction (having to call customer service and wait for an hour), inertia (being on the same Movistar plan for years), or decisions we never revisit (my grandpa paying for life insurance since he was 30).
There is a long tail of these problems. Each one might be too small or too infrequent to justify the effort. You know you could find a cheaper plan or dispute that charge, but you have other things to do. So it stays there, sometimes for years.
AI changes that calculation because agents are infinitely patient. They can compare options, check the details, follow up, and keep going through the boring parts that make us give up. They also give people access to help with decisions that previously required expertise they didn’t have or couldn’t afford.
Something that wasn’t worth an hour of your time might be worth asking an agent to handle. And while each problem is small, there are a lot of them. Each problem might only come up once a year, but across all of them, there could be something worth solving every week.
Now, the challenge is the same one we’ve been talking about over and over. AI can only help if people recognize the opportunity and know it can help.
There are two ways this may happen: pull or push.
Pull means people experimenting and figuring out the power of AI for these problems. You ask it to review your Movistar plan, find a better option, and help you switch. There is a chance that, with widespread adoption, enough viral “this AI solved that problem why don’t you try it” examples will get people to try it, and then come back with another problem.
One example I keep seeing go viral is organizing school information for parents. Javi and I worked on this a year ago, and it is available in Luziclaw. (Though there is a cheaper solution: schools should get their shit together and learn to communicate concisely.)
Push means AI that is integrated enough into the consumer’s life to be proactive. Proactivity is the next big thing. An assistant with access to your bills could notice that your Movistar plan deserves a review and suggest a better option before you ask.
That is still a work in progress. We are getting better, but there is a gap between “organize my email and tell me something” and actually resolving the problems it finds. The opportunity is to make that long tail worth dealing with, without turning the user into the full-time manager of another tool.
As AI starts solving more of these problems, there is a whole other topic here: what happens to industries when people stop staying on old payment plans, pursue every valid insurance claim, or ask their bank to refund every incorrect charge.
But that’s for another post.
