In 2022, Kasey Hayano began remodelling her home. It had been her grandmother’s, and Hayano was keen to do the job properly, but the 1960s building proved trickier to handle than she had initially imagined. “As we were remodelling,” she recalls over Zoom from her home in California, “we quickly realized that it was a lot more complex than you think it is [going to be] when you go into it.”
Coupled with the budgetary restrictions she felt as a new homeowner, Hayano began to wonder if there wasn’t an easier way to manage and streamline her project. She’d been working in Bay Area tech start-ups for a decade, and – with the characteristic problem-solving approach of a long-time Silicon Valley ops manager – started mulling over potential solutions, even deciding to change careers to work at a Palo Alto interior design firm to crib some professional experience. “I wanted to get hands-on design experience, because I recognise that I don’t have a formal interior design background. I spent days, weeks emailing all the local interior designers in my area asking them if they needed any help.”
The career switch paid off: after working at a studio for several months, Hayano identified rendering – the process of creating a 3D digital visualisation of a design scheme – as one of the most time- and energy-consuming tasks that her colleagues faced, and decided to use AI to streamline and automate the process. In February last year, she co-founded HomeVisualizer, a suite of AI tools based on the StableDiffusion model of generative AI technology (GenAI), which takes image inputs and turns them into detailed renders for homes. She is now the CEO.
A HomeVisualizer user can upload a picture of a room in their or their clients’ home, as well as an “inspiration” image of another room, and then (alongside a written description of the desired elements of the inspiration image), the AI will generate an uncannily accurate render that applies the aesthetic of the inspiration to the original picture, but with elements such as furniture and fittings left in situ. Other options include automatically applying a pre-defined style – such as “Coastal”, “Industrial” or “Art Deco” – to a photo of a room. The idea is that this tool can help designers and clients to visualize their home were they to take their design down different paths.
Instagram content
Traditionally, Hayano explains, taking time to render (as in the amazing image above, created by the Egyptian-American designer Hassan Ragab) can be expensive, and is often unavailable to those renovating their own homes without a designer’s help. So while Home Visualizer might be just one tool in the arsenal of an interior designer, making rendering quicker and easier, it can also act as an off-the-shelf product for amateur designers who want to imagine their own remodels quickly and easily. When I uploaded a photo of the poky, slightly worse-for-wear shared bathroom of the ex-council flat I live in into Home Visualizer, for example, and asked the tool to reimagine it in a “Japandi” style, it took about 45 seconds to return a render of a calming new bathroom replete with wooden tops and a soft grey colour palette.
HomeVisualizer is one of many new tools – often, but not always, using GenAI – to supplement more traditional design practices. In many cases, these tools aren’t intended to replace a designer at all, but shave time and hassle off elements of the home renovation or buying process.
For example, one online tool called REimagineHome claims to offer virtual staging for estate agents, with the option to add or remove furniture from photos of houses, and to redesign landscaping around a building (House & Garden tried this one out, with mixed results – as with everything AI, many of these tools are both very new and totally unregulated, so quality can vary). Other design-adjacent tools span image editing software with AI-powered features (Canva, Playground.com) and room-specific platforms (LivK, for your kitchen. These tools feel playful to use; occasionally, it’s hard not to think of using them as something akin to gaming, perhaps designing an interiors scheme in a much more advanced version of The Sims.
Instagram content
Entrepreneur Jenna Gaidusek, who founded of AI for Interior Designers in October 2023, is one of the industry’s most prominent exponents for using AI. After graduating formally with a degree in interior design in 2010, Gaidusek became interested in virtual design while working for a furniture company, and in 2015 she set up her first business to pursue this new, digital design avenue full-time. In 2018, she founded a Facebook group called eDesign Tribe; the community would eventually gain some 5,000 members, and developed into an online school for designers called eDesign U, whose use cases increased almost exponentially during the Covid-19 pandemic, as designers scrambled to learn how to work with clients remotely and virtually. “It took off like wildfire,” recalls Gaidusek, “because everybody needed to transition.”
Like Hayano, Gaidusek identified AI and rendering as an area of potential synergy (Gaidusek is now quoted prominently on the HomeVisualizer website, and the two know each other professionally). She is a vocal champion of using AI in the industry; since selling her platform to another design project-management system called Mydoma in 2022, she now leads AI classes and webinars, including teaching clients how to use MidJourney and ChatGPT.
Gaidusek is keen to emphasise that AI is not a prescriptive tool, but an instrument to be used by people – like a pencil, or Photoshop. Nor is it producing anything other than a very general idea of what a space might look like, in a very general aesthetic style. “None of these images are actually scaled – you can put in measurements and [MidJourney] ignores them. We’re not using actual products. They’re purely conceptual.”
Hayano agrees: “We get this question all the time: ‘Can you put this specific rug in this specific room?’ We’re like, ‘No – it’s not there yet.’” She says that GenAI is more akin to a mood board than a detailed plan for a design scheme. Hayano and Gaidusek each make separate, positive comparisons to Pinterest when explaining the current state of interior design AI, and Gaidusek estimates that she’s generated “probably tens of thousands” of images using AI over the last few months. And yet, she says, “even if you have the ideas, retrofitting them into your space still needs a designer.”
So – the consensus is that AI isn’t going to come and eat interior designers’ lunch, at least for now. And yet, there’s a hint of just how it might change design practices, when you talk to those involved in its cutting edge. “Right now, we're just getting started,” says Gaidusek. “It’s so hard to see where we’re going to be in six months, let alone five years – because it is moving so rapidly.”
