Façading
Façading refers to the practice of presenting a product or company outwardly as its own AI development, even though internally foreign technology or even human labor is actually handling the task. The term is mainly used in reports about overvalued start-ups and misleading marketing.
Façading comes from the French word for facade, i.e. the front of a house. It refers to a company that looks like its own technology developer on the outside, but uses foreign work on the inside. For example, a provider sells a supposedly proprietary chat assistant. In fact, its program simply forwards every request to a large external provider and returns that provider’s answer. The company’s own contribution then consists only of a surface, a logo, and billing. In extreme cases, humans even sit in the background typing the answers that are presented as machine work.
Why this matters
For investors, it makes a huge difference whether a company owns its own technology or merely resells someone else’s. Proprietary technology is hard to copy and justifies high valuations. A pure facade can be rebuilt within a few weeks. That is precisely why façading is treated as a warning sign in financial news.
On top of that comes a cost risk. Anyone who merely resells is dependent on their supplier’s price. If the supplier raises its fees, the business model collapses. There have been several prominent cases in which start-ups lost almost their entire value after such revelations. For customers, data protection is also affected, since their inputs end up with a third party they know nothing about.
How it works
Technically, façading is simple. Almost all major AI providers offer an interface, i.e. an address on the internet to which a program can send queries. A few lines of code are enough to address this interface. The provider behind the facade then builds its own app with its own design.
The user sees none of this. They type their question into an attractive interface and receive an answer. The fact that the computing work takes place in a foreign data center remains hidden. Some companies add small proprietary building blocks of their own, such as text templates or a search function for company documents. In that case, the line between honest reuse and deception becomes blurred.
What matters, therefore, is not the technology but the communication. Anyone who openly states that they are building on foreign models is conducting normal business. Anyone who speaks of “our own AI” while merely forwarding requests is engaging in façading.
Where the term is encountered
Façading has most frequently appeared in investigative reports about start-ups that raised large sums with supposedly proprietary technology. A well-known pattern: an app promises automated programming, but in the background developers in another country are actually doing the work. Such stories appear regularly in business media.
It’s also worth taking a closer look in everyday life. Many apps sold as AI tools for homework, image editing, or translation are facades in front of well-known models. Often the actual provider is listed in the terms of use, buried far down in the fine print. If such an app costs more as a subscription than the original, you are mainly paying for the surface.
The topic is also addressed in reports by auditors and regulatory authorities. In the EU, advertising regulations require that statements about products must not be misleading. Exaggerated AI claims can therefore have legal consequences. For investors, a simple rule of thumb applies: ask what the company has built itself and what it merely buys in.