Am I obsolete?

I was working on a keynote this weekend, packaging up what I had learned from two decades of obsessing over users - and building winning products in the process.

The results speak for themselves - Redbox, one of the fastest companies to $1B in revenue (at the time). RetailMeNot, a unicorn IPO. H-E-B, rated the top grocer in the country based in part on the strength of the digital experience.

But that question kept popping in my head - am I obsolete? When AI can replace the user, does obsessing over them even matter?

What I believe: the technology is new, but the human experience in relation to it is not.

I've built a bunch of products that were either agentic - like paying someone else to do your grocery shopping for you - or predicated on extreme convenience and time savings.

That's the bedrock of AI value - have another entity do the thing for you, and save boatloads of time.

Some old truths, and new ones, for building inspired products today:

1) Just add an egg.

When Betty Crocker launched their cake mix in the 1950s, it was a flop. All you had to do was add water, which was too convenient. But change the recipe and require the user to add an egg, and success! The user could now claim emotional ownership of the end result.

For AI products, we need to add intentional friction to create emotional ownership.

2) Save time, make 💰

Humans are addicted to convenience. From cake mix, to hiring out your grocery shopping, to AI agents. All we want to do is do less. We're not awed by what the technology can do, we're awed by how much we don't have to do.

Humans will pay absurd amounts to claw back time. Just study your next grocery or DoorDash delivery bill and see what you really paid for the convenience, vs the food.

3) Respect human preferences.

Ah, but where do agentic experiences go horribly awry? When the agent doesn't do it just like we would do it! If I have a hankering for homemade guacamole (yes, I know that's slower than buying it premade), place an order, and rock hard avocados show up 😠.

That's just a simple example. Truly exceptional agents need to understand their user's preferences at incredible depth - like they're living in their brain.

4) Humans expect perfection from machines.

For our entire lives, computers have been deterministic - put in a set of inputs, and it will reliably spit out the exact same thing, every time.

That era is gone. AI is probabilistic. It can do super complex things, but the output is a highly educated guess - which means it can and will get things wrong. We haven't fully wrapped our brains around this, and it's going to take a while before we do.

5) Acquire users - or replace them.

This is the brand new one for the AI era, especially for B2B. There may, in fact, not be a user. Just a buyer, and what we're selling the buyer on is - "you no longer need all these humans doing [insert X task]."

(I still haven't figured out the long term implications of this one)