CHATTERgo agents act on your store data and talk to your shoppers, so the platform is built around a few firm security rules.
Grounded answers
Agents answer from your synced catalog, orders, and knowledge base — not from their own invention. They do not reveal their internal instructions, model details, or other shoppers' data, and conversational guardrails block attempts to extract such information.
Guardrails without false positives
Guardrail patterns are deliberately tuned so ordinary shopper language is never caught: first-person order questions ("list my orders"), shipping questions in any language, and product names that happen to contain technical words all pass through. Only genuinely suspicious patterns — third-party data probing, prompt-injection style requests — are blocked, and repeated abuse can earn a temporary session ban rather than an immediate one.
Identity precedence
A shopper's verified storefront identity (established directly with Shopify) always outranks any generically exposed browser data, so order and account information is only shown to the person actually signed in. Guests are asked to sign in for identity-gated actions rather than being trusted on a browser hint. See Customer identity.
Tenant isolation
CHATTERgo is multi-tenant: your catalog configuration, knowledge base, conversations, and analytics are scoped to your organization. Platform-level connections — such as the Store Agent's read-only store access or the ClickUp integration — run brokered and scoped to what the connected account is allowed to do.
Access control for your team
Organization members sign in with their own accounts. You can define custom roles with a permission matrix, control which agents members can see and manage, and review changes in an audit log.