AI copilot for homeownership
SaaS that ingests homeowner records so AI can triage repairs, answer maintenance questions, and broker context-rich proposals to contractors.

- Team
- Solo-Founder (Designer, Engineer, Sales & Support)
- Stack
- Next.js (web), Expo + React Native (mobile)
- Audience
- Homeowners & Home Services
- Timeline
- Mar 2024 — Sep 2024
- Scope
- SaaS spanning a consumer-side home-knowledge operating system (ingestion, AI Q&A, repair triage) and a B2B contractor-side reverse-proposal marketplace, connected by an AI assessment layer.
- Responsibility
- Everything: product strategy, design system, frontend, backend, infra, marketing, sales, design, support.
Context
What do you actually know about your home? Most homeowners hold the answer in a folder of PDFs, a stack of invoices, and the memory of which contractor showed up last time the water heater leaked. And contractors get inbound from people who can't articulate the problem and don't have the history to share—so quotes are blind, trust is low, and the easy work goes to whoever happens to be available. humbleOS is the layer between the two: an AI home operating system that holds the homeowner's records, answers their questions, and brokers context-rich work to the contractors who can actually estimate it.
Approach
Two surfaces, one knowledge layer. On the homeowner side, ingestion is the contract—inspection reports, invoices, appliance details, photos and video become structured context the AI can reason against. On the contractor side, reverse-proposals flip the usual flow: homeowners post a need with the full record attached, and contractors counter-offer based on real context instead of a 30-second phone description. The AI sits in between—triaging the issue, estimating the repair before anyone gets called, and deciding whether a contractor is even needed.
Impact
Shipped two pilot projects with contractors and homeowners. The load-bearing lesson: the contractor business is hyperlocal—different licensing, pricing dynamics, and trust networks in every region—and hyperlocal doesn't venture-scale cleanly. A well-designed product isn't always a venture-backed business, and humbleOS turned out to be one of those.

How do you design for contractors without being one?
To design for contractors, I took a project management job and worked alongside them—the kind of unscalable research a founder has to do before there's anything to scale. I watched how they negotiate with homeowners, build budgets, send proposals end-to-end, and where the workflow actually breaks. The product is built on what couldn't be learned from a survey.

What does the homeowner side look like?
You feed it everything—inspection reports, invoices, appliance manuals, photos of what's where—and ask anything. Look up local building code for a permit, draft a renovation plan, find out which repairs are tax-deductible, check state or federal compliance. The mountains of one-off research a homeowner usually has to do collapse into one AI surface with the home's full context already attached.

How do reverse-proposals work?
Pros receive detailed asks from homeowners with the issue pre-assessed by AI. Review on the go, dictate a response, and send a custom proposal in minutes—no callbacks, no trip back to the backoffice to file a new proposal in a separate system. The contractor who responds with real context wins the work before the next bid even lands.
More work
AI-native WYSIWYG builderInstacartAug 2025—May 2026
Live music discovery for the obsessedHumbleMar 2025—Present
Organizing unstructured work dataPanoramaSep 2024 — Nov 2024
Scaling IG & FB Shop internationallyMetaMar 2020—Nov 2022
Design tools for 1,000+ designersMetaJun 2018—Mar 2020
Consumer IoT experience + design systemRingJul 2016—Jun 2018