The latter is reasonable. The former isn't. You should be able to come up with lots of ideas worth following up on.
Let's play a game. Someone give me a product category, and I'll list ideas that are at least worth more research.
- Made to order furniture. Let users design their own, or at least mix and match high level componets at a finer detail than ikea.
- High end IKEA. Combine the flexibility of Ikea with high end designs and materials.
- High level office furnishing service. Companies provide a high level description of their furnishing problem, rather than choosing specific items. I'm sure these consultants already exist, but maybe it can be done in a more automated, scalable way.
- bid/ask interface for used furniture markets. This would enable sellers to passively sell items.
- used marketplace that uses ML to extract features from pics. I think this may already exist. Not sure.
Very few pieces are composable, we need more things like https://www.prettypegs.com/ Unless you have the capability to build a new IKEA, this is to me the best venue for providing high quality customizability.
Finding furniture is a mess. There's missing information (dimensions, max load), close to useless search, and a bunch of places that sell the same things.