AI Home Design That Goes All the Way to the Building Permit

01 /The Concept

Drafted is an AI platform where anyone can design and generate a complete architectural plan for their future home.

The process starts with the conceptual layout: users select rooms from the platform's catalog by function, square footage, and style. The AI then generates five full home design options. The user picks the closest match, shares a list of preferences, and the system generates five more. This iterative loop continues until the result is satisfying.

At that point, payment kicks in – and the output is a complete home project delivered as a PDF or a set of AutoCAD files. It includes facade plans, floor plans, roof plans, electrical layouts, and even a construction schedule and materials delivery timeline. Pricing scales with complexity but runs between $1,000 and $2,000.

The resulting plans can go straight to a construction firm for cost estimation. From there, the builder can refine the project with the client or proceed with the delivered design as-is.

Drafted launched this year and has just raised its first funding round: $1.65M on a $35M valuation.

02 /Why It Matters

The founder of Drafted isn't new to this space. In 2020, he launched Atmos – an online company built to deliver custom home designs faster and cheaper than traditional architecture firms. He went through Y Combinator that same year with Atmos, grew it to $7M in annual revenue, and raised $20M along the way.

But as he later admitted, Atmos had become "a purely operational business" – essentially a traditional architecture firm dressed up in technology. That bothered him enough that nine months ago he shut Atmos down.

He then founded Drafted on the same underlying theme – and has since been accepted into Y Combinator's next spring batch. This time, his goal is to build a pure technology company: an AI engine for generating home designs, not another service operation.

With this positioning, he's aiming to land between two extremes – slow, expensive architecture firms on one end, and cheap but poorly customizable off-the-shelf home plans on the other.

The founder's trajectory is instructive because it reflects a dilemma that confronts every tech entrepreneur who steps into a traditional industry.

At some point, founders discover they're standing at a fork in the road:

- Path one: keep building what is essentially the same traditional business, propped up by technology. This inevitably becomes a conventional operational business with all its tradeoffs.

- Path two: radically simplify the process. This usually produces a cheap but inflexible product – one that only appeals to buyers who can't afford something better, or who find the quality acceptable.

For a long time, the gap between these two extremes has been largely unfilled. The occasional attempt to bridge it – either premium providers launching cheaper products, or cheap providers trying to sell more flexibility and quality – rarely succeeded. The two business models don't coexist gracefully; trying to stretch in either direction corrupts the existing processes and creates internal contradictions.

The only credible solution now visible is the AI-native product. These products offer far greater flexibility than low-end alternatives while remaining genuinely affordable – landing somewhere in the middle between the extremes.

The real question is whether the middle is wide enough to build a large business on. The conventional wisdom is that you make serious money by selling expensive and scarce, or cheap and abundant.

Zooming in on the home design market: of the roughly 1.5 million homes built in the US each year, only about 300,000 are built to custom designs.

One of Drafted's investors frames this as a chicken-and-egg problem. Cheap custom designs fail to satisfy; proper ones are too expensive for most buyers, or psychologically hard to justify spending serious money on.

His view: if high-quality custom designs become affordable and fast enough to produce, demand will materialize. He draws the analogy to Uber – before which the taxi market served a fraction of today's riders, at far lower frequency.

Whether that logic holds in home design remains to be seen. But the underlying principle has merit. The closest precedent is Shopify – whose founder couldn't raise money early on because investors thought 40,000 online stores wasn't a big enough market for a venture-backed startup.

He believed that removing friction – making it dramatically easier and cheaper to start and run an online store – would grow the market itself. He was right. Shopify now hosts over 1.5 million stores.

03 /Opportunities

Conservative service markets are among the most exciting opportunities for technology startups right now – precisely because meaningful breakthroughs are now achievable through intelligent use of AI. This is a real trend, and a growing number of startups are acting on it.

But founders who enter these markets will eventually face the same fork – build a traditional business, or retreat into the cheap-and-narrow niche. Neither path is inherently wrong, but not every founder wants to run an operational business, and that's a character question as much as a strategic one.

The interesting direction, then, is the AI-native product that can serve as the "golden middle" on a market that matters. That's the path Drafted's founder is now walking.

What conservative service market seems large enough and interesting enough to you? What AI-native product could land in the middle between the expensive/slow and the cheap/inflexible? What would it actually need to do?

Is there a chicken-and-egg problem on that market – insufficient demand for the middle option today? Could that demand grow substantially if someone offered acceptable quality, speed, and price?

For what it's worth, that chicken-and-egg problem should be present – because if it isn't, the market is already fully served. And if it is, there's a real opening.