No. The system prompt frames it as an assistant for the coaching business, not as the coach. It can describe the coach's methodology, certifications, and packages, but it will not roleplay as the head coach during a sales conversation. Prospects who ask programming-specific questions get routed to the discovery call. That separation protects the coach-client relationship and avoids the awkward moment when a prospect realises they were talking to a bot for 20 minutes.

Yes, conversational intake usually outperforms a long form for online coaches. SleekAI walks the prospect through goal, training history, equipment access, schedule, and injury notes in plain language, then emails the structured data to the coach. Most coaches see a 40 to 60 percent higher completion rate compared to a 25-field form. Real Typeform integrations can still run alongside if you want a backup capture.

No, and the system prompt explicitly refuses to. Programming is the coach's job. The bot can explain the methodology (block periodisation, conjugate, RP-style hypertrophy, polarised endurance) at a marketing level, but it will not prescribe sets, reps, percentages, or weekly volume. That refusal is non-negotiable for liability and for the coach-client relationship.

The bot is configured to recognise ED history (current or past) and to route the conversation to a coach review with a gentle, non-clinical reply. It will not give nutrition advice, will not propose a calorie target, and will not run a body-composition discussion. For active ED concerns the bot points to NEDA (1-800-931-2237 in the US) or an equivalent local hotline and lets the coach decide whether the engagement is appropriate.

It can hand off to any of them. SleekAI does not write into TrueCoach, Trainerize, or My PT Hub directly. It captures the discovery call booking, the goal, the training history, and the intake answers, then pushes the lead via webhook or email so the coach can onboard inside their existing platform. The bot is the front of the funnel; the coaching platform is the workspace.

The system prompt forbids body-composition timelines, weight-loss promises, and outcome guarantees. Most over-promising in fitness chatbots comes from a permissive prompt, not from the model. SleekAI's prompt explicitly bans phrases like "lose 20 pounds in 8 weeks" and routes any direct outcome question to a discovery call with the coach. The guideline filter catches the same pattern even if the LLM drifts.

Yes. The bot can explain that the coach uses block periodisation with two heavy days and a metcon, that nutrition is habit-based not macro-based, or that the program rotates 4-week blocks. That is methodology, not prescription. It will not write "Monday: squat 5x5 at 75%". The line is enforced in the system prompt and reinforced by the guideline filter.

Yes. SleekAI is bring-your-own-key, so the prompt goes from your WordPress server directly to OpenAI, Anthropic, Google, or OpenRouter. Conversation logs live in wp_postmeta on your site, not in a SaaS dashboard. The coach owns the data the same way they own the WordPress install. Set retention to match your privacy policy and document the chatbot in your intake disclosure.