IKEA: a beautiful wardrobe still has to fit the room

There is a moment in a furniture project when an attractive visual meets the radiator below the window. The radiator usually wins. It is a useful metaphor for AI in retail: impressive imagery is easy; preventing a real purchasing mistake is harder. Especially when that mistake arrives in several heavy boxes.

IKEA in the US describes a bedroom measuring and planning service involving measurements and the room's requirements. That is an established service. My scenario below is an independent proposal for preparing digitally for that kind of work, not a description of an existing IKEA AI deployment.

Home photographs contain personal details, so storage, deletion and a dimensions-only option must be clear. Marketing use requires separate permission. The furniture AI I want does not promise a perfect home. It helps ensure a chosen wardrobe does not require the customer's actual home to urgently reconsider its geometry.

Begin with constraints

The first useful input is a set of verified dimensions: walls, ceiling height, openings, obstructions and door movement. AI explains what still needs measuring through clear language and a simple diagram. It should not confidently extract exact centimetres from an arbitrary wide-angle photograph. A lens can make a room more generous than an installer will find it.

These dimensions allow conflicts between a chosen product and the space to be checked. Does the cabinet fit, can its doors open, is the passage usable, and is assembly possible? Hard constraints are calculated separately from generated prose. An unknown measurement should pause the recommendation, not disappear beneath an inspiring description of future comfort.

The collected information can then become a concise consultation pack. Customers share confirmed measurements, uncertainties and alternatives; a specialist does not need to collect everything again. AI prepares the draft and a person checks it. This preserves the human conversation for decisions that genuinely require experience and taste rather than repeated administration.

Returns begin before payment

Restricting a trial to one furniture system and straightforward room types keeps the result interpretable. One group receives the measurement assistant; a comparable group follows the ordinary instructions. Assess what happens to orders rather than the attractiveness of the pictures. Complex installations go directly to a specialist. Someone else's walls should not fund experimental confidence.

My three measures would be dimension-related returns, repeated clarification before ordering and consultation time to an agreed plan. Conversion is secondary here. Good advice sometimes prevents an unsuitable purchase. The retailer loses that transaction but avoids a round-trip delivery and an understandably irritated person holding a tape measure.