How the AI generates emails

When you click Generate Draft, here's what happens behind the scenes.

The generation pipeline

1. Gather context
   - Product code, name, market, description
   - Lead details (company, contact, segment, notes)
   - Lead's pain cards (if any)
   - Knowledge Base search (relevant docs)
   - Writing style profile (if configured)

2. Build prompt
   - All context assembled into an instruction

3. Generate
   - LLM creates a personalized email draft

4. Post-process
   - Check quality (length, personalization)
   - Apply A/B test variant (if active)
   - Save as pending draft

Why emails are personalized

The AI doesn't just swap names. It uses multiple layers:

Layer	Example
Role-based	Different tone for CTO vs Sales Manager
Company-based	References their industry, size
Pain-based	Addresses their specific problems
Context-based	Pulls from Knowledge Base
Style-based	Matches your writing patterns

AI providers used

Outlix uses different models for different tasks:

Task	What's used
Email drafts	Fast LLM (Groq/Llama)
Pain extraction	Fast LLM (Groq/Llama)
Niche documents	Claude Opus (deep analysis)
Document search	OpenAI embeddings

You never interact with these directly — the platform routes automatically based on the task.

Human-in-the-loop

The AI generates. You decide.

Every email is a draft until you approve it. The AI never sends anything on its own. This is by design:

Prevents spam
Protects your reputation
Ensures quality
Keeps you in control