Account & Onboarding Guidance
How does multi-tenant organization isolation work?
When you create an account, AIXONICS provisions an isolated organization record. All connected channels, Brand Brain documents, drafts, and execution logs carry your unique organization ID. Users can belong to multiple organizations and switch contexts without cross-tenant data leakage.
What roles are available and who can authorize publications?
AIXONICS enforces granular Role-Based Access Control (RBAC): Owner (full governance and billing), Admin (workspace and channel management), Manager (publishing approval and team oversight), Approver (authorized to approve post dispatches), Editor (can create and adapt drafts), and Viewer (read-only audit access). Only Owners, Admins, and Managers/Approvers can authorize external dispatches.
What happens if our organization reaches its monthly AI token budget?
When your monthly spend reaches 100% of the configured limit, the AI Gateway engages a hard-stop barrier. Generative completions are temporarily paused to protect against budget overruns until an Owner or Admin raises the ceiling in Organization Settings.
Social Channel Integration Troubleshooting
LinkedIn: Why did my account connection show "Token Expired"?
LinkedIn OAuth tokens require periodic refresh. While AIXONICS runs background token refresh jobs before expiration, if an operator changes their LinkedIn password or revokes third-party app permissions directly on LinkedIn, the token must be re-authenticated by clicking "Reconnect" in Settings > Connected Channels.
Meta Facebook/Instagram: Why do Instagram posts fail without an image?
The Meta Instagram Content Publishing API strictly requires a publicly accessible media container (image or video) for every broadcast. AIXONICS automatically validates this before allowing submission, preventing broken dispatches.
X (Twitter): Why was my tweet rejected for character limits?
X enforces a strict 280-character ceiling for standard accounts. AIXONICS provides a live character countdown in the composer. If you have links or attachments, the character budget adjusts automatically to match X URL wrapping standards.
Email Account & Delivery Troubleshooting
Microsoft 365: "Need Admin Approval" error during Graph connection?
If your Microsoft 365 tenant restricts user consent for third-party applications, your Azure AD / Entra ID Global Administrator must grant admin consent for the AIXONICS application scopes (`Mail.ReadWrite`, `Mail.Send`, `offline_access`).
Gmail: Connecting with custom App Passwords vs. OAuth2?
We strongly recommend connecting Gmail via our official Google OAuth2 integration. If you are using Google Workspace with custom SMTP credentials, you must generate an App Password in your Google Account security settings and enable IMAP.
Why was an email draft held in "Pending Approval" status?
To prevent accidental dispatches and comply with anti-spam policies, AI-assisted email replies can be configured to require managerial approval before actual SMTP or Graph delivery occurs.
Content Generation & Brand Brain Troubleshooting
Why did the AI output omit a specific product detail?
Generative prompts retrieve context from your Brand Brain vector store using semantic similarity. Ensure the relevant document has been uploaded and successfully chunked in the Brand Brain tab, and that relevant keywords are included in your prompt topic.
Can I use my own OpenAI, Anthropic, or Gemini API keys?
Yes. AIXONICS supports full Bring-Your-Own-Key (BYOK) architecture. In Settings > AI Gateway, supply your organization keys. Credentials are encrypted at rest with AES-256 vault encryption.
What is the canonical microcent accounting precision?
To ensure accurate token metering across multi-user teams, AIXONICS calculates all model costs down to $10⁻⁸ USD integer units. This eliminates floating-point rounding errors and matches upstream API provider billing exactly.
Autonomous Agents & Safety Guardrails
Why did an Agent execution pause with status "AWAITING_APPROVAL"?
This is an intentional security guardrail. Whenever an autonomous agent attempts an action that broadcasts to public social channels or client inboxes, the deterministic engine pauses execution and requests human verification.
What is Loop Protection and why is it active?
Loop Protection is an automated safeguard that limits agent execution steps (default: 15 steps per run). If an agent becomes caught in a repetitive tool query or reasoning cycle, the execution engine halts the run to conserve your AI token budget.