Why no code tools for building ai models failed?

Business folks have tasks where neural networks could help but they don’t know Python or deep learning, right? Meanwhile ML engineers know AI models but often don’t have an actual problem to solve. So the first group hires the second.

Same thing happened with websites a while ago—at first you had to code everything then came drag-and-drop editors, no-code tools, and now even this "vibe coding" stuff. It’s not the most flexible or maintainable, but it works for a lot of people who just need to get something done without spending months learning to code.

So why isn’t the same thing happening with building AI models?

Like if I run a business I could save a ton of money by automating manual tasks with models trained on my own data. But I have zero clue what Jupyter Notebook or Keras are, and honestly, I don’t want to know. So I have to hire an ML engineer just to get it done because there’s no viable no-code solution for this.