AI Disclosure

You are talking to AI, not a human

Slí is operated by Lugha — Learn Upskill Grow Harness AI Limited. Slí's coaching assistant is powered by large language models through a provider abstraction. Providers may include Anthropic, OpenAI, or another configured model provider depending on the tenant and deployment. The platform identifies the assistant as AI, never as a human coach.

What the AI sees during a session

When you have a coaching conversation, the AI may receive:

• The current message you sent
• The recent message history of this specific session
• Memory entries or profile context you have explicitly consented to keep across sessions
• The active stream framework and competency dimensions configured by your coach or organisation
• Recent learning or reflection signals when they are explicitly bridged into a session

What the AI does not see

• Other users' conversations or memory
• Your messages to a human coach in non-AI messaging channels
• Your account credentials or payment information
• Data from another tenant or organisation

Your data is not used to train models

Slí sends relevant conversation context to configured model providers for inference only. Your conversations, reflections, and memory are not used to train underlying AI models through Slí's configured provider path.

Boundary detection and safety

Coaching conversations run through boundary-detection rules that look for crisis, harm, medical, legal, or other out-of-scope situations. When a boundary is detected, Slí redirects toward human support or an appropriate external resource rather than letting the AI continue as if it were qualified to advise.

Memory and deletion controls

Memory is designed to be under participant control. You can review, export, or delete data through Profile → Legal & your data. For the wider data-rights story, see the Privacy Policy.

AI errors and recourse

AI output can be wrong, incomplete, or context-blind. Treat it as coaching support, not as professional advice. If the assistant says something incorrect, harmful, or out of scope, you can report it in the product or raise it with your human coach or organisation administrator.