> ## Documentation Index
> Fetch the complete documentation index at: https://docs.bitscale.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# AP-7: AI Personalization Capstone

> Build and run a complete AI-personalized outbound campaign from scratch. 50+ contacts, multi-layer personalization, full QA system, live sends. Pass and earn the AI Personalization Specialist certification.

<Info>
  **AP-7 · AI Personalization · 300 XP · \~3–4 hours build + 1 week running**
</Info>

This is the capstone for the AI Personalization track. You're building a fully personalized outbound campaign — multi-layer personalization, QA infrastructure, live sends, results tracking.

***

## What You're Building

A production-ready personalized outbound campaign with:

1. **Prospect grid** — 50+ contacts from a defined ICP
2. **Multi-layer personalization** — company signal + role challenge + individual hook
3. **Full prompt system** — with vertical-specific variants
4. **QA infrastructure** — all 4 checks + master status + auto-retry
5. **Live sends** — at least 40 approved emails sent via sequencer
6. **Results tracking** — reply sentiment, personalization level correlation with reply rate
7. **Capstone analysis** — did personalization depth correlate with reply rate?

***

## Build Instructions

**Step 1: Source and enrich (45 min)**
50+ contacts with full enrichment: firmographics, funding, tech stack, signal, LinkedIn research.

**Step 2: Multi-layer personalization (60 min)**
All three layers for every contact:

* Company context synthesis column
* Role challenge column
* Individual observation column
* Personalization classification (full / two-layer / etc.)

**Step 3: Prompt system (30 min)**
Build prompt variants:

* Full personalization prompt (Layer 1 + 2 + 3)
* Two-layer prompt (Layer 1 + 2, no individual signal)
* Company-only prompt (Layer 1 only)
* Vertical-specific prompt (for your primary vertical)

**Step 4: Generate and QA (30 min)**

* Run all 4 QA checks
* Master QA status
* Auto-retry for failures
* Manual review of flagged rows

**Step 5: Send (ongoing over 1 week)**

* Export approved rows to sequencer
* Send minimum 40 emails
* Track replies in a results grid

**Step 6: Analyze (30 min)**

* Tag each reply with sentiment
* Calculate reply rate by personalization level
* Did higher personalization = higher reply rate?

***

## Capstone Hypothesis

Every submission must include a hypothesis test:

*"Contacts with full three-layer personalization will have a higher reply rate than contacts with two-layer or company-only personalization."*

Test this against your actual data. Document:

* Reply rate by personalization level (full / two-layer / company-only)
* Whether the hypothesis was confirmed, partially confirmed, or rejected
* What this suggests for your future workflow

***

## Certification: AI Personalization Specialist

Pass this capstone to earn the **AI Personalization Specialist** certification.

Requirements for passing:

* 40+ emails sent
* Reply rate > 5%
* QA infrastructure fully implemented
* Hypothesis test documented with real data

<Card title="Submit AP-7 Capstone →" icon="trophy" href="https://bitscale.fillout.com/academy-challenge-ap7">
  Submit your grid + results + hypothesis analysis. **+300 XP on approval. Certification unlocked.**
</Card>

***

<Card title="Next Track: GTM Architect →" icon="arrow-right" href="/academy/gtm-architect/overview">
  You've mastered data, signals, and personalization. GTM Architect is where you learn to design and build complete GTM systems from scratch.
</Card>