What happens when AI becomes the interface for everything?

OpenAI had their annual Dev Day last week – a showcase of what’s next in generative AI. And they dropped big news: apps are moving inside ChatGPT.

This doesn’t mean integrations or plugins. We're talking apps running natively, inside the AI. Built with an SDK that’s apparently so simple, you just need to export HTML.

It’s a bit mind-bending. The norm is to put AI into apps. Now, the script is flipped. Let’s unpack why, and what it means for business and for users.

So what?

If ChatGPT essentially becomes the ‘interface’ for the internet and digital world – which is clearly the direction we’re moving in – then businesses need to pay attention.

Initial launch partners like Canva, Spotify, and Zillow are already live. Later this year, developers will be able to submit their own apps for approval; more or less the same way you’d get a new app onto the app store.

Here’s the kicker: if your app isn’t in ChatGPT, and your competitor’s is, you just became pretty much invisible. Because users won’t be browsing websites anymore. They’ll be prompting ChatGPT, because it’s easier, more intuitive, and becoming ubiquitous. ChatGPT will recommend the app within its own ecosystem that can fulfill their request.

Ask it to build a playlist? It’ll suggest Spotify. Want to turn a poster into a pitch deck? Canva’s baked in. Looking for homes in Pittsburgh? Zillow’s got you covered. And in the normal flow of LLM conversation, you can simply ask things like how close the nearest dog park is, or to find properties with solar panels – you name it! The context flows back and forth between the app and the AI.

What about existing AI tools?

Many popular tools are designed to bring AI into the software people use daily. Everything seems to have its own AI assistant built in now, right? But ChatGPT is doing the reverse. It’s bringing the tools into the AI.

So which way wins? Do we keep enhancing traditional apps with AI? Or do we start designing for a future where the AI is the primary access point – and apps are just capabilities that plug into it?

Prompting is the new UX

In a world where AI becomes the main way we interact with the digital world and array of tools in it, prompting skill suddenly matters a lot more – and it’s already critical!

The prompt doesn’t just shape the output – it determines which agent within OpenAI gets activated to deal with your task. A vague prompt might trigger a misplaced agent less capable for the output you really need. A nuanced prompt with context, signaling the type of operations the agent will need to perform, unlocks powerful targeted functionality.

For example, the difference between a ‘creative content’ agent vs an ‘analytics’ agent can take your workflow way off-piste.

In terms of consumer awareness, we’re lagging. There’s still confusion around what an “agent” actually is. In OpenAI’s world, an agent needs access to tooling and autonomy to qualify. But in enterprise settings and society, we often use “agent” to mean any AI assistant. The gap in understanding could lead to misaligned expectations – and missed opportunities.

The AI Divide

My broader concern is that as soon as OpenAI effectively replaces our traditional ways of accessing digital products and services, those who are behind on AI literacy are at a big disadvantage as consumers.

The knock-on problems are tricky to predict, but it’s broadly fair to say that the ‘AI Divide’ (a phrase I might be coining here) is going to be worse than the ‘Digital Divide’ – where those without home internet access or technical skills have been at a massive disadvantage over the last 25 years.

AI for Everyone

We talk a lot about ‘AI skills for everyone’ in the training world, and the odd thing is that the ‘everyone’ in question can often be taken to mean ‘non-technical’ people – there’s a perceived split between the AI makers and the AI takers. Builders vs users. It’s not that simple.

In reality, we’re all AI end users now. AI will be helping the techies to build the tech, just as much as it’s helping the non-techies to boost their efficiency and utilize tech.

Whether you’re a developer, a product owner, knowledge worker or a business leader, the shift to AI-native interfaces means rethinking how your services are discovered, accessed, and experienced by everyone.

The same goes for businesses. There is a potentially radical change in the consumer landscape coming that you must anticipate. And the bottom line seems to be: if your app isn’t inside the AI, it might as well not exist.