Is OpenAI Building the Real Super-App?

The Western world has chased the idea of a Superapp for many years - trying to recreate Wechat in the west, the Chinese social media, messaging, and wallet app owned by Tencent. Within Wechat are many "mini-programs" which exists from Didi (cars) to bike sharing to paying for utilities.

Yesterday, the CTO of Instacart announced they are partnering with OpenAI on Apps within ChatGPT which looked interesting. One of the more interesting screenshots (we will see in a few months the usefulness) I've seen in a while. But will it cook?

What's a SuperApp?

The corporate vision is to be the toll collector for everything - more expansive than even the mobile app stores. Many companies in the United States have professed to be building super-apps. From X to Square/Block and more. Even Klarna and Bolt have mentioned the idea in the last year. Most is just marketing hype.

But no one has the potential to pull it off like OpenAI. The success of OpenAI becoming a superapp depends on a few things:

1 - Incentives for the developers. The "vig" from OpenAI needs to be not so onerous as to discourage innovation. Apps must be easily promotable and the app store must be seamlessly integrated with the experience. (read: Ads are eventually coming here too)

2 - User ease. This is a big one. The flow has to be natural for users, such that it doesn't require special code words or awkward things you type to activate the usefulness of the app. Think about the Alexa app store as the worst example of this. Technically it works, but practically it's painful. LLMs have the promise of making it easy.

3 - Network effects. More users attract more app developers. More app developers build more apps. More apps means greater chance of useful apps. Useful apps attract more users.

What would kill such a super app opportunity?

- It's too clunky. Language is imprecise, which could make it hard to use or find apps.

- Early app failures. Big early wins could help fuel the effort. Big notable failures could cool off the effort.

- Competition. Someone like Apple or Google who already has apps. Are they thinking this way? Hard to tell. Counting out Apple never a good idea. They are one of the best fast-followers in the world when a big new idea hits the stage and is clunky as hell to use. Few have the user obsession that Apple has. To OpenAI's credit, they do recognize this. But with Jonny Ive really working on hardware, will it get much of his time?

Lots of questions, few answers at this stage. But the ambition makes it one to watch.