Define automated quality checks that run on every message before it's submitted. Each check can use multiple AI models and flag messages with specific concerns for the reviewer to address.
{{variable_name}} syntax in your prompt templates to insert these values directly. All fields are also sent as structured JSON context to the AI — you don't need to use variables if you reference the field names in your prompt instructions. New rules should generally just check "Full Research Archive" below and use {{full_research_data}} rather than hunting through the individual fields further down.{{full_research_data}} — Everything used to generate this message, formatted as one clearly-labeled block: full LinkedIn profile, every post fetched, the specific post replied to (if any), full news article, About Me Connect findings, and the sender BDR's full background/instructions. No truncation, no picking individual items — just check the Full Research Archive field below and use this one variable.message_generation_archive (see about_me_message_lookup.html to view it per-message). Falls back gracefully to whatever's on the message doc for records older than the archive or past its retention window.
{{message}} — The outreach message text (after any edits){{message_type}} — "connect" or "message"{{source}} — Source type (LinkedIn Post, Org News, etc.){{first_name}} — Contact's first name{{last_name}} — Contact's last name{{full_name}} — Full name combined{{title}} — Job title{{company}} — Company name{{linkedin_url}} — LinkedIn profile URL{{post_text}} — Full text of the LinkedIn post (always included){{post_url}} — URL of the post{{post_type}} — regular, repost, or quote{{post_date}} — ISO date the post was published (e.g. 2025-11-03){{post_age}} — Human-readable age: "3 weeks ago", "5 months ago", etc.{{reshared_post_text}} — Original post text for reposts/quotes{{reshared_post_author}} — Author of the original post being reshared{{news_data}} — Internet search news stories (JSON){{ai_company_news}} — AI-summarized company news{{ai_contact_news}} — AI-summarized contact news{{ai_fit_reason}} — AI-generated fit explanation{{profile_headline}} — LinkedIn headline{{profile_about}} — About / summary section{{profile_experience}} — Work experience entries{{profile_education}} — Education history{{profile_connections}} — Connection count{{about_me_findings}} — About Me Connect analysis data{{original_ai_message}} — Pre-tone-review AI message{{bdr_name}} — BDR's name (sender){{source_batch_label}} — Batch label if applicable{{bdr_profile}} — BDR's sales config, value prop, approach, and "I" statements from About MeDownload all rules as a JSON file to back them up. (Rules already sync automatically across browsers via the shared store.)
Import rules from a JSON file. Choose whether to replace existing rules or merge.
Remove all review rules from the shared store (affects every browser and generation page). This cannot be undone (unless you export first).