AI Governance
Last updated:
Lala is FoxtINN's AI virtual manager (distinct from Laila / Call Laila for demos, and from Leela / Voice AI). This document describes the governance, data handling, and safety practices we apply to Lala. White Glove is a separate human managed service and is not an AI product.
Grounded responses
Lala is designed to ground answers in your live operational data and to cite sources (for example shift, location, or ticket) where the product surfaces citations. Where citation-backed data is unavailable, Lala is instructed to refuse or ask clarifying questions rather than invent facts.
Refusal behavior
When Lala cannot answer with sufficient grounding, it should refuse rather than guess. Refusals may be logged for product improvement within your tenant controls.
Training data
Your operational data is not used to train public foundation models. Improvements are limited to your tenant or, with explicit consent, aggregated and de-identified forms.
Bilingual evaluation
Lala is evaluated in English and Spanish across hospitality, retail, restaurants, and related scenarios. Evaluation results may be shared with customers under NDA on request. Product UI and staff-facing content support EN and ES; this website is primarily English.
Human-in-the-loop
Actions with material operational impact (for example shift offers or vendor work orders) require human approval by default. Auto-approval is opt-in per workflow where offered.
Auditability
Lala actions intended for the audit trail are captured with relevant prompt context, retrieved context, response, and approving user where applicable.
Questions? Call Laila at (816) 816-3823 · hello@foxtcon.com · Back to Trust Center