Post SMTP keeps every send in a custom log table with subject, recipient, status, mailer, and the raw response. The current transport, OAuth tokens, sender details, and backup configuration live in the postman_options option, with state in postman_state_options. The plugin also writes to its own debug log and supports Chrome push and mobile app notifications for failures. Most of that signal never reaches a support agent.

SleekAI maps those fields as prompt variables. The bot can tell a visitor that the active transport is Microsoft 365 OAuth, that the token refreshed three hours ago, and that the last failure was a 535 on the backup SMTP because the app password is stale. Display conditions scope each bot by URL pattern, role, or post type, and conversations log model name, token count, and origin URL.

Generic bots send users to the wrong settings panel because they do not know Post SMTP from PHPMailer or WP Mail SMTP. They miss the difference between SMTP-API and OAuth transports, they ignore the mobile app integration, and they cannot explain why an email failed yesterday because they cannot read the log. A SleekAI bot can.

Post SMTP has been around since 2017 and now serves more than 400,000 active sites. It supports OAuth for Gmail and Microsoft 365, classic SMTP, several API mailers, fallback connections, and a notification system that includes Chrome push and a dedicated mobile app. That capability surface is exactly where generic chatbots fall over.

A generic bot does not know whether you are on OAuth or SMTP. It has no idea your token refreshed three hours ago. It cannot tell you that the last 14 sends failed with InvalidAuthenticationToken because it cannot read the log table.

So it sends users to wrong settings panels and burns trust. SleekAI fixes this with simple variable mapping. The bot reads the log, the option, and the state, and the system prompt teaches it the language of Post SMTP.

The result is short, accurate answers like reconnect the Microsoft 365 app or turn on the Brevo fallback, with a real reason behind each one. Privacy stays under your control. The model call uses your own API key, the data the model sees is only what you expose, and the guideline filter blocks the bot from leaking credentials.

Token usage and model name are logged per conversation so you can prove cost against ticket deflection before you scale to more sites.