AI applications in the last few years have been called "ChatGPT wrappers" as a way to undermine their value.

The argument was that OpenAI, Anthropic, or any other foundational AI company could easily win their market share because AI application companies were nothing more than pretty-looking frontend wrappers on foundational models. For a while, this was true. Scrappy AI products built in a few days or weeks were just ChatGPT didn't do much beyond give GPT-provided answers.

But even a year later, there's a very real argument that the AI segment most ripe for disruption is the application layer.

If each AI application is a wrapper, then delightful packaging has proven to be a differentiator. OpenAI's GPT Store, which many believed would dominate the AI application layer, flopped.

Meanwhile, the most widely praised and adopted AI products are standalone apps. As Chris Pedregal from Granola, an AI meeting notepad, writes:

"Exceptional experiences for narrow use cases often have little to do with AI. We spend endless hours on note quality at Granola, but we spend just as much time on features like seamless meeting notifications and great echo cancellation (so our tool works whether you're using headphones or not). The "wrapper" around the AI is often the difference between a delightful experience and a great demo that is disappointing to actually use."

This is true beyond AI as well. Many of the best consumer products could be called "API wrappers". Innovative technical products are great, but they're not the only way to solve a user's pain point. More often than not, pain points are centered around UX.

---

Find more of my thoughts on Twitter (@ethanweii).