Hint's home AI assistant: the asset-record pattern

Hint, a home management app co-founded by Martha Stewart, launched on iOS on July 29 with $10 million behind it. The press coverage is about the celebrity. The interesting part is the architecture: Hint's AI assistant is the last layer, not the first one, and the pattern underneath it is one we'd copy for almost any business that services physical assets.

What actually happened

Per TechCrunch, Hint is co-founded by Stewart with CEO Yih-Han Ma and CTO Kyle Rush, backed by $10 million from Montauk Capital, Slow Ventures, Tusk Venture Partners, Energy Impact Partners, Amplo VC, Hannah Grey, and Brian Kelly of The Points Guy. Stewart holds equity and works on the product rather than writing a check.

The flow: you enter your address. Hint assembles a profile of the house from public data — property records, weather, soil information, utilities — before you've typed anything else. Then you attach documents: inspection reports, warranties, mortgage paperwork, insurance policies, invoices, photos of appliances. From that it generates a maintenance schedule with push notifications and a "home score." The chatbot, built on OpenAI and Gemini models, answers questions against that record.

The app is free on iOS at launch with no ads or subscription. Revenue comes from affiliate deals with service providers, which the company says is firewalled from what the AI recommends. Paid tiers — multiple properties, for one — are planned.

Why the asset-record pattern matters for your business

Four moves, in this order, and only the last one is a model:

1. Seed the record from public data. Nobody fills out a blank form. County assessor data, permit history, parcel records, and utility territory are public and mostly free — enough to render a populated profile before the customer does any work. This is the single highest-leverage thing in the whole design and it has nothing to do with AI.
2. Attach the documents. Warranties, service history, spec sheets, photos. This is the moat: it's the data no competitor can reconstruct and the reason churn hurts.
3. Derive obligations. A maintenance schedule is a query over the record plus a calendar. Deterministic, testable, and the thing that actually creates repeat contact.
4. Then put a model on top, scoped to answer from the record rather than from the open web.

Swap "house" for "HVAC fleet," "rental portfolio," "commercial kitchen," or "shop equipment" and the build is identical. If you run field service, property management, or equipment sales, you already hold steps 2 and 3 in a filing cabinet and someone's head. Turning that into a queryable record is a normal software project — a Postgres schema, a document store, a scheduler, and a retrieval layer — not a $10M venture bet.

One detail worth stealing and one worth interrogating: the affiliate firewall is the right instinct, because the moment recommendations follow commissions the assistant is an ad. Ask any vendor selling you this pattern how that separation is enforced in code, and who audits it.

Sources: TechCrunch, Inc..