Style transfer is the umbrella for any AI technique that takes "what this looks like" from one source and applies it to "what this is" in another. In interior design, it's the core operation: your room is the content, the chosen style (Japandi, Mediterranean, Coastal) is the style.
Modern style transfer in design tools is implemented through a combination of img2img (preserves the room's structure) and either prompt-based conditioning (the prompt names the style) or reference-image conditioning (you upload a photo of the look you want, the tool matches it). The reference-image variant — sometimes called "image prompting" or IP-Adapter — is more controllable than text alone because the model has a concrete visual target to match.
The practical advice: if you can find a reference photo that exactly captures the look you want, use it as input. The output will track that reference more reliably than any prompt could.