Runnr.ai

Delivers pre-trained natural language understanding specifically optimized for hospitality guest inquiries (room service, housekeeping, check-in/out, amenities, billing) rather than generic chatbot responses. The system uses domain-specific intent classification and response templates trained on hospitality conversation patterns, enabling accurate handling of context-specific requests without requiring extensive customization by property staff.

Automatically detects guest message language and responds in the same language without requiring explicit language selection, supporting multiple languages simultaneously across a single chatbot instance. Uses language identification models (likely fastText or similar) to classify incoming text, then routes to language-specific response templates or translation pipelines, enabling properties to serve international guests without hiring multilingual staff.

Operates continuously without human intervention, automatically classifying incoming guest messages by complexity and routing simple inquiries to pre-trained responses while escalating complex issues (complaints, special requests, emergencies) to appropriate staff members with full conversation context. Uses intent confidence thresholds and rule-based routing logic to determine escalation paths, maintaining conversation history for seamless handoff to human agents.

Allows properties to customize pre-trained hospitality responses with property-specific information (amenities, policies, contact procedures, branding) through a configuration interface without requiring code changes or model retraining. Uses template substitution and rule-based customization to inject property data into responses while maintaining consistency with hospitality best practices and tone.

Aggregates and analyzes guest conversations to identify common inquiry patterns, frequently asked questions, and guest satisfaction signals without requiring manual log review. Generates reports on inquiry types, response effectiveness, escalation rates, and language distribution to help properties optimize staffing and identify gaps in pre-trained responses. Uses basic NLP metrics (intent distribution, response acceptance rates) and statistical aggregation.

Connects to property management systems (PMS) via webhooks or APIs to access real-time property data (occupancy, guest profiles, maintenance status) and trigger staff notifications (SMS, email, push) when escalation is needed. Enables context-aware responses (e.g., 'Your room will be ready at 3 PM') and ensures escalated issues reach appropriate staff immediately rather than sitting in a queue.

Maintains conversation history across multiple guest messages, enabling the chatbot to understand references to previous messages ('Can you repeat that?', 'What about the WiFi?') and provide coherent multi-turn responses without losing context. Uses conversation state management to track guest intent across turns and avoid repetitive responses, improving perceived intelligence and guest satisfaction.

Offers free tier with limited conversation volume, languages, and customization depth to enable small properties to test the platform, with paid tiers unlocking higher limits and advanced features. Implements usage tracking and quota enforcement to manage free tier costs while providing clear upgrade paths for growing properties. Likely uses API rate limiting and feature flags to enforce tier restrictions.