Straight answers, including about what isn't built yet.
An AI Operations Platform that connects a business's own knowledge, its live operational context (reservations, orders, tickets), real conversations, and a controlled layer of automation and human escalation — built to work across property, ecommerce, customer support, and ISP/NOC operations on one shared architecture.
No. A chatbot answers questions. AnswerBnB resolves who and what a conversation is about first, retrieves only knowledge scoped to that context, and chooses between answering, clarifying, escalating, or flagging a conflict — a chat interface is only the visible surface of that.
Every conversation is tied to a real organization and, where applicable, a specific property, reservation, order, or ticket at the data level — this is enforced in the application and database, not inferred loosely from message text alone.
Each property or service can have multiple knowledge sources — manual entries and crawled URLs. Sources are fetched respecting robots.txt, extracted into text, deduplicated by content hash, and stored as chunks the AI can retrieve from — scoped to the property that owns them.
It says so and escalates to a human with the full conversation attached, instead of guessing. Confidently answering the wrong thing is treated as a worse outcome than an honest "I don't know."
Detecting conflicting knowledge (for example, one source saying check-in is 3 PM and another saying 4 PM) and surfacing it rather than silently picking one is part of the Knowledge Engine's designed architecture, being built out on top of the current single-source-per-fact foundation.
Yes — escalation to a human is a core part of the design, with full conversation and grounding context carried over, not a cold handoff.
That's the intent of the architecture: a normalized adapter layer designed to sit around property-management, ecommerce, CRM, help-desk, and ISP/NOC systems rather than requiring you to replace them. See Integrations for exactly what's built today versus designed for the future.
Systems like Guesty, Hostaway, Hospitable, Lodgify, OwnerRez, and Smoobu represent the kind of PMS ecosystem AnswerBnB's architecture is designed to connect with.
Platforms like Shopify, WooCommerce, and BigCommerce represent the kind of ecommerce ecosystem the architecture is designed to connect with.
Systems like Salesforce, HubSpot, Zendesk, and Intercom represent the kind of CRM/help-desk ecosystem AnswerBnB is designed to sit around.
Systems like Splynx, Sonar, UISP, and MikroTik-based networks represent the kind of ISP/NOC ecosystem the architecture is designed to connect with.
No. Every vendor name is labeled honestly — Available, In Development, Architecture Compatible, or Planned — and as of today, no third-party vendor connector is live. Naming a platform describes the ecosystem we're building toward, not a claim that a live connector exists.
Sources track when they were last crawled and last verified, and content is hashed so an unchanged re-crawl doesn't create duplicate knowledge. A fuller freshness/staleness status model is part of the Knowledge Engine's ongoing build-out.
Yes, and doing so is designed to be able to feed back into the organization's approved knowledge — so the same gap doesn't need correcting twice — rather than only affecting the single conversation.
The platform's automation is built around a strict principle: it may only execute actions it has been explicitly authorized to take. A general-purpose automation/journey engine (trigger → condition → AI decision → action) is architecture and roadmap today, not a shipped feature.