An interior is lived in, not only looked at. Its light, sound, surfaces and layout shape ordinary routines long after the presentation is over. In this seventeenth essay in my 2024 research series, I explore how AI could expand the design process while keeping that everyday experience at its centre.

Generated images can help a designer and client discuss atmosphere, compare styles or try unfamiliar combinations. They are proposals, not specifications. An apparently convincing room may contain impossible joins, inaccurate dimensions or products that do not exist. Surveying the space, drawing it properly and checking materials remain essential.

Dark sculptural interior with an indoor tree · Concept illustration
Dark sculptural interior with an indoor tree · Concept illustration

Personalisation begins with listening. Preferences change with time, budget, ability and the people sharing a home. A system may find patterns in choices, but it cannot infer every need from them. A useful interior leaves room for habits, possessions and adjustments that were not anticipated in the initial brief.

Bespoke furniture and flexible zoning could make small or complicated spaces work better. AI may support the exploration of options, but dimensions, ergonomics, construction and access still need verification. Adaptability can be as simple as a movable partition or replaceable component; it does not require a constantly transforming, automated room.

Brick-walled loft with large windows · Concept illustration
Brick-walled loft with large windows · Concept illustration

Material and lighting choices must work beyond the screen. Samples, maintenance requirements, glare, acoustics and performance in actual conditions can change a decision. Environmental value depends on durability, repair, sourcing and replacement as well as appearance. A convincing visual mood does not prove comfort or sustainability.

Virtual walkthroughs and clear visual comparisons can improve conversations with clients, whether or not they use AI. Procurement tools may assist with schedules, availability and alternatives, but cannot guarantee flawless delivery or installation. Changes should be checked against the agreed design, budget and the realities of the site.

Connected controls might adjust lighting or other services to selected settings. They should be understandable, optional where possible and usable manually. Learning a routine is not understanding an occupant’s feelings. Collecting less personal information and allowing people to change settings themselves can matter more than elaborate automation.

The cultural question is whether easier image-making leads to more personal spaces or more versions of the same fashionable room. I am interested in interiors that can acquire character through use. The most valuable role for AI may be to help explore possibilities while leaving the final place open to the people who live there.