Staynup logo

Why does the AI need to know my property?

Staynup's AI is a powerful language model, but it is not omniscient. When a guest messages you, the AI doesn't reach into some shared pool of knowledge about your home — it works from what you have told it. This page explains where that knowledge comes from and why feeding it well is the single biggest lever for automation.

The principle

The AI knows nothing about your property until you give it knowledge. It can only answer a question using facts you have provided — your listing details, your rules, and the corrections you teach it. It cannot and will not invent details on its own.

This is by design. If the AI could guess, it would eventually guess wrong — telling a guest there is a pool when there isn't, quoting the wrong wifi password, or promising a late checkout your cleaner can't support. Grounding every answer in facts you control keeps the AI accurate, consistent, and safe to leave unsupervised.

The practical upshot: the quality of your automation is limited by the quality of your knowledge. Every fact you add makes the AI slightly more useful. Every missing fact is a question the AI must hand to your team instead.

Three sources of knowledge

There are exactly three ways knowledge gets into the AI. Together they build a complete picture of how you run your property.

Knowledge sourceYour Airbnb listing
Where it comes fromThe listing details you already filled in on Airbnb
How you feed itSynced automatically when you connect your listing
Knowledge sourceCustom rules
Where it comes fromPolicies and instructions you write yourself
How you feed itTyped in the dashboard, or taught by messaging the manager channel
Knowledge sourceCorrections via chat
Where it comes fromFixes you make when you see the AI get something wrong
How you feed itMessage the manager channel in plain language, e.g. "change the wifi password rule"

The first source is the foundation — most properties get 80% of the way with a complete listing alone. The second and third sources are where you tune the AI to your specific way of working.

Knowledge is automation

Automation quality equals data quality. Think of the AI's routing this way: for every guest message, it checks whether it can answer from facts you have given it. If yes, it replies automatically. If no, it hands off to your team.

  • A complete listing plus a few good rules means the AI answers around 95% of routine guest questions alone.
  • A thin listing means constant handoffs — every trivial question lands in your team's lap, and you're back to answering messages yourself.

The good news is that you don't need to get this perfect on day one. You can build knowledge gradually: fill in the essentials, watch what the AI hands off, and add the missing facts as you go. Each addition permanently reduces future handoffs.

Tip:

Most hosts reach strong automation by completing the Automation Checklist — every item adds knowledge the AI can use.