Legal
AI & Data Practices
Last updated: August 2026
Your data does not train AI models. Audit context processed by our AI features is used to generate a response for your facility, and for nothing else.
1. Where AI Is Used
- Coach PC — real-time observation guidance while a supervisor conducts an audit.
- Coaching points — talking points generated from a completed audit's failed criteria.
- Training plans — in-service lesson plans generated from the facility's recent audit patterns.
2. What the Model Sees
AI requests include only the minimum context needed for the feature: the audit criterion being evaluated, the lifeguard's first name and recent failure patterns at your facility, the zone, and the certifying body's standards. Requests are scoped to a single facility — the model never receives another client's data in your requests.
3. No Training on Client Data
AI processing is performed via the Anthropic API. Under its commercial terms, API inputs and outputs are not used to train Anthropic's models. PoolControl.ai does not train, fine-tune, or build models on client data, and does not permit any subprocessor to do so.
4. Human in the Loop
AI output in the Service is advisory. Audit scores are determined by the supervisor's own pass/fail judgments — never by the model. Generated coaching points are editable by the supervisor before delivery, and generated training plans are reviewed by management before use.
5. No Cross-Client Exposure
Facility isolation applies to AI features exactly as it does to the rest of the Service: the context for your requests is drawn only from your facility's records, and generated output is stored only in your facility's account.
6. Contact
Questions about AI practices: poolcontrolnate@gmail.com