Questions, answered plainly

No. The AI reads a request and recommends a department, priority and action from the rules your hotel has configured. A person approves the task, a person decides whether to escalate, and a person sends any reply to a guest. These controls are built in and cannot be switched off.

Every recommendation shows the matched terms, the rule it came from and a confidence score. Low-confidence or unmatched requests are flagged “needs human triage” instead of being routed. Safety-related terms (gas, fire, medical and similar) always map to an urgent, manager-alerting rule, whatever the model says.

The design uses Amazon Bedrock for classification and summaries, so you can pick a model your AWS account has access to. In this prototype Bedrock is not connected: the demo runs on a deterministic rules engine, and the Bedrock code path is written but untested against a live account.

Yes. Each rule defines keywords, category, department, priority, SLA, recommended action and whether to suggest escalation. Managers can edit them in the dashboard and every change is written to an audit log. The AI can only choose among the rules you enable.

Not yet. Requests currently come from a web form or are logged by the front desk. Messaging is planned through an authorised provider, with guest opt-in and approved templates. Nothing in this prototype sends messages to guests.

The same engine applies, with apartment-specific rules such as appliance faults, mid-stay linen changes, utilities queries and long-stay recurring issues. Properties can be hotels or apartment buildings and are managed together.

Yes. Staff see only the properties and departments their role allows. Group managers get cross-property views and can compare recurring issues across sites.

The target architecture stores tasks, rules and property data in PostgreSQL on AWS, and guest attachments in Amazon S3 with encryption and malware scanning. The prototype itself keeps sample data in your browser and stores nothing on a server.

No certifications are claimed. The product is at prototype stage. The architecture is designed around data minimisation, role-based access and audit logging, but formal compliance work has not been done.

Not yet. PMS integration for room status, guest profiles and VIP flags is planned. VIP flags and rooms in the demo are sample values.

No. The pricing page shows placeholder launch pricing for the prototype and will change before general availability.

It is a working prototype with fictional sample data. The website, dashboard, rules engine, handover and analytics work end to end in the browser. AWS services, HubSpot and guest messaging are designed and partly scaffolded but not connected.