By default SleekAI refuses to give medical advice and points users to a healthcare professional. The wording is configurable in your prompt - some stores route to their in-house naturopath or pharmacist, some to a generic 'consult your doctor or registered dietitian' line. Medication interactions, condition-specific dosing, pregnancy and lactation questions all hit the same wall. The bot is friendly about the deferral.

Yes. Servings, dosage per serving, and ingredient amounts from your product meta or labels feed into the answers. A shopper asking 'what's in your magnesium glycinate?' gets the actual mg per capsule. The bot does not extrapolate to 'take 2 capsules at night for sleep' - it quotes the label's directions and notes that personal dosing is between the customer and their doctor.

Upload your compliance and warning pages into the knowledge base and the bot stays inside that language. DSHEA-compliant structure-function claims, mandatory warnings (pregnancy, allergens, age), and third-party testing references all stay inside your approved copy. The bot does not promote a product as treating, curing, or preventing a disease.

Yes. Multibot lets each line have its own prompt and display conditions. The sports bot on /sports-nutrition/ knows protein, creatine, pre-workout categories. The general bot on /vitamins/ knows multi-vitamins, mineral chelates, herbal extracts. Each has its own tone and its own guardrails calibrated to the category.

No. Display conditions scope each bot to specific URLs, categories, or post types. A shopper on /sports-nutrition/ does not get a herbal-extract answer; a shopper on /sleep/ does not get a creatine recommendation. The bot stays in its lane because the prompt and knowledge base are scoped to that lane.

Yours. SleekAI is bring-your-own-key, so usage and costs stay on your account. Compliance-sensitive categories also benefit from data-residency control - conversation logs hit your WordPress database, OpenAI calls run on your account, and there is no third-party platform layer that could log conversations to a separate vendor.

Same guardrail as other medical questions - 'I can share product info but I'm not a doctor; for pregnancy or breastfeeding questions please consult your healthcare provider.' Pregnancy categorization on supplements is heavily regulated and varies by ingredient, so the bot does not improvise. The conversation gets logged but does not trigger a marketing follow-up.

Yes. If your products carry certification flags in product meta and your knowledge base includes a 'what these certifications mean' page, the bot explains them neutrally. NSF Certified for Sport, USP Verified, and GMP-compliant manufacturing each mean different things, and customers shopping for athlete-safe or third-party-tested products want a clear answer instead of marketing-speak.