A First Job to be Done for AI

That Apple is having a hard time integrating AI is not too much of a surprise. At the heart of their challenge lies a Core Design Dilemma: ow to use AI for the benefit of the everyday user expereince whereas the current breed of AIs are targetted for text-prompt-driven and technology-ladened “creative” uses? Turns out, a little application of Jobs To Be Done can be a great guide in integrating AI into everyday user experiences.

Lost in Transportation

On a recent trip to NYC, I encountered a frustrating multitasking UX mishap. Yet, in it was a lesson as to where AI can fit into our everyday lives.

I used the Maps app to get directions to my next appointment, which was in 30 minutes. The app indicated a leisurely 24-minute walk to midtown. This seemed quite doable as I walk briskly and can usually beat the Map app’s estimates. As I hit the button on my phone, I received a call from my mother. No problem, I thought: I can walk and talk. Twenty minutes later, I looked at my phone and realized I was still 23 minutes away! How could this be?

It turned out my Maps app’s default was set to driving mode. I had inadvertently followed a long path that was best for driving, not walking. As it turns out, in NYC, driving and walking paths are quite different and can be difficult to tell apart on a map.

One could call this an operator error or blame poor map labeling, but such transportation mode decisions likely happen billions of times every day. The extra cognitive load of deciding which mode to use distracts from more important tasks such as where am I going, especially in real-time situations like driving, biking, or crossing the street, where distractions can be costly.

Better yet, why is there even a default mode for transportation? This choice is a complex, multi-faceted challenge, a perfect use case for the supercomputer we happen to carry in our pockets.

Learning from Struggling Moments

This UX anomaly in the otherwise excellent Apple Maps app got me thinking: perhaps this type of trivial problem is exactly where AI can get its foothold with the end users: as a humble, silent assistant who handles simple, thankless tasks that can easily be corrected, such as selecting the default mode of transportation based on my specific context, location, and geography. Surely, the smartphone that counts my steps can easily detect that I was walking, not driving. Why didn’t it alert me that I was using the wrong mode? Better yet, the Maps app can learn from my corrections and improve its future recommendations.

Entry-Level Jobs

The technology world’s chatter today is all about grand visions of AI as an all-encompassing mastermind capable of making complex, high-level decisions. Yet, the true promise of AI today lies not in constructing grand, idealized, or dystopian futures. Instead, it lies in subtly enhancing our current, ordinary lives, empowering us to overcome everyday Struggling Moments.

This approach positions AI as a supportive tool—an assistant that doesn’t replace, but enhances, our experiences by freeing us from the trivial and tiresome. What we need today is an AI eager to start its work at the entry-level of human actions, and not one intent on carrying out a hostile executive-level takeover.

An AI-integrated smartphone is uniquely positioned to serve billions of users in trillions of mundane situations. Here, we can find a great application for AI in the background, utilizing those idle computing moments, and many opportunities for AI to learn from its mistakes.

Updated: 25 Nov 2025

---

Related Lenses