ParkView AI

Small guest houses receive complex group-booking enquiries through WhatsApp. Guest numbers change, couples need shared beds, elderly guests need accessible rooms, and getting the accommodation and quotation right takes hours. ParkView AI is a working hospitality prototype built around one principle: Meta AI reads. Code decides. The owner approves. Meta Muse Spark interprets natural-language enquiries, including Afrikaans, corrections and relative dates. Its Evidence Trail identifies the guest's exact words supporting extracted booking facts. Deterministic code checks room capacity, bed arrangements, accessibility, availability and pricing. Unresolved questions prevent premature approval, and the owner retains final control. The prototype uses fictional guest-house data and clearly distinguishes live Meta API calls, recorded replays and simulations. We documented 20 genuine Meta API calls, including Muse Spark and Muse Voice Transcribe, and completed 272 automated tests. Built as a new codebase during the hackathon, ParkView AI demonstrates how language AI and verifiable business rules can work together to reduce booking mistakes for small accommodation businesses.