Crm journey architect

CRM Journey Architect

Design CRM journeys that are operationally executable on SFMC or AJO, data-aware, consent-compliant, measurable, and connected to business outcomes. Treat CRM as last-mile delivery: the journey must move the right customer, through the right trigger, to the right action, with the right message, at the right time.

How This Skill Orchestrates

1. Web search (Claude — MANDATORY): Search for current email/CRM benchmarks by [INDUSTRY] [COUNTRY] [LIFECYCLE STAGE]. See references/platform_implementation.md → "Mandatory Benchmark Search Protocol"
2. Script execution (scripts/journey_builder.py): Generates structured journey spec from lifecycle stage + segments + triggers + channels, validates data requirements, scores readiness, produces platform-aware implementation checklist (SFMC or AJO)
3. Reference lookup (references/platform_implementation.md): SFMC/AJO implementation guides, journey pattern library (5 pre-built patterns), CRM benchmarks by industry (2025-2026 verified), data architecture requirements
4. Upstream context (Claude): Pull audience segments from audience-segmentation-brief, channel allocation from media-routing-planner, scope from scope-audit
5. Visual output (Visualizer): Journey flow diagram + readiness scorecard + data dependency matrix

Model decision: Entity extraction + document intelligence (Scope layer). Journeys are conditional logic trees — not predictions. No ML needed. The script generates the journey structure; Claude adds the strategic layer (messaging hierarchy, offer strategy, exception handling). The 70/30 split: ~70% of CRM operations are AI-automatable (template builds, segmentation, trigger config, QA, scheduling, reporting), ~30% requires human judgment (strategy, creative, offers, compliance, exceptions).

The user MUST name the CRM platform (SFMC or AJO). The implementation checklist, data model, personalization syntax, and common failure modes are platform-specific. If not stated, ask before proceeding.

Quick Reference

Resource	Purpose	Usage
scripts/journey_builder.py	Journey structure generator — lifecycle stages (11 types), platform-aware steps (SFMC/AJO), data requirement validation, readiness scoring, 70/30 task split	python journey_builder.py --input config.json --output journey.json
references/platform_implementation.md	SFMC 12-step implementation checklist + AMPscript patterns, AJO 12-step checklist + streaming/batch guidance, 5 journey patterns (welcome, abandoned cart, winback, post-purchase, lead nurture B2B), CRM benchmarks 2025-2026 by industry, data architecture requirements, LATAM adjustments	Read for platform-specific implementation and benchmarks

How to Use the Script

import sys
sys.path.insert(0, "<skill-path>/scripts")
from journey_builder import JourneyBuilder

config = {
    "client_name": "AcmeAuto MX",
    "journey_name": "Aftersales NPS Journey",
    "lifecycle_stage": "aftersales",
    "platform": "sfmc",
    "business_objective": "Increase NPS score and drive service appointments",
    "segments": [
        {"name": "Recent buyers (0-30 days)", "definition": "purchase_date >= today - 30", "size": "~2,000/month"},
    ],
    "triggers": ["purchase_event"],
    "channels": ["email", "sms", "whatsapp"],
    "data_fields": [
        {"field": "customer_id", "source": "Salesforce CRM", "available": True},
        {"field": "email", "source": "Salesforce CRM", "available": True},
        {"field": "consent_status", "source": "Consent DB", "available": True},
        {"field": "vehicle_model", "source": "DMS", "available": True},
        {"field": "purchase_date", "source": "DMS", "available": True},
    ],
}

builder = JourneyBuilder(config)
result = builder.build()
# result.steps, result.data_requirements, result.readiness_score, result.platform_setup_checklist

Process

Step 1 — Define Business Outcome

What commercial result does this journey drive? Not "send emails" but "increase 90-day repurchase rate by 15%."

Step 2 — Define Lifecycle Stage

Map to one of 11 stages: acquisition, onboarding, activation, consideration, conversion, repeat purchase, loyalty, cross-sell, upsell, retention, winback, aftersales. Each has different goals, channels, and benchmarks.

Step 3 — Define Segments

Who enters this journey? Define segments with data criteria, estimated size, and priority. Use audience-segmentation-brief if available.

Step 4 — Define Triggers

What event starts the journey? Form submission, purchase, inactivity, lead score change, app event, service event, renewal window. Map triggers to the platform's entry source capabilities (SFMC: DE/API event, AJO: segment/business event).

Step 5 — Design Channel Sequence

Which channels, in what order, with what timing? Email → wait → decision (engaged?) → SMS if no → final email. Use journey patterns from the reference as starting templates.

Step 6 — Define Message Logic

For each message: value proposition, CTA, personalization fields, fallback content, timing relative to trigger.

Step 7 — Map Data Requirements

What data fields are needed? Customer ID, email, consent, lifecycle stage, plus stage-specific fields. Run the builder to validate.

Step 8 — Define Suppression & Governance

Consent status, frequency caps, unsubscribe handling, cooldown periods between journeys, regulatory compliance (GDPR, LFPDPPP for Mexico, LGPD for Brazil).

Step 9 — Define Measurement

KPIs per step (open rate, CTR, CVR), journey-level KPIs (completion rate, revenue attributed), benchmarks from searched data.

Step 10 — Produce Implementation-Ready Journey

Generate the full spec with platform-specific implementation checklist.

Output Format

Produce in THREE forms: journey flow diagram (Visualizer), readiness scorecard (Visualizer), and structured markdown spec.

Journey Flow Diagram (Primary Visual)

Render the journey as an inline HTML widget showing a vertical step flow:

• Each step as a card: type badge (entry/message/wait/decision/exit), channel icon, timing, CTA
• Decision splits shown as branching paths (Yes/No)
• Color coding: green (entry/exit), teal (message), gray (wait), coral (decision)
• Data requirements flagged on each step
• sendPrompt() buttons: "Run QA for this journey" → campaign-launch-qa, "Show SFMC setup checklist" → loads platform reference

Readiness Scorecard (Secondary Visual)

Render as an inline card:

• Readiness score (0-100%) with status badge
• Data field status (available / missing / unknown)
• Content status per step
• Platform setup checklist with check marks
• Blockers list

Markdown Spec (Tertiary)

## 🔄 CRM JOURNEY — [Journey Name] — [Client]

### Objective
[Business outcome the journey drives]

### Platform
[SFMC / AJO] — [specific implementation notes]

### Audience
| Segment | Definition | Size | Priority |

### Journey map
| Step | Type | Trigger/Condition | Channel | Message role | Timing | CTA | Exit |

### Data requirements
| Field | Purpose | Source | Required | Available | Risk if missing |

### Content requirements
| Asset | Channel | Personalization | Owner | Status |

### Suppression & governance
[Consent, frequency caps, cooldown, compliance]

### Measurement
| KPI | Definition | Benchmark ([source]) | Frequency |

### Platform setup checklist
[SFMC or AJO specific — from reference]

### 70/30 task split
| AI-automatable (70%) | Human-required (30%) |

### Readiness
Score: [X]% — [status]
Blockers: [list]

### Recommendation
[What to build first, what to test, what to defer]

Examples

Example 1 — Abandoned cart (SFMC): User: "Build an abandoned cart flow on SFMC for our e-commerce client in Mexico. Average cart value $85." → 3-step journey: Email (1h after abandon) → SMS if no open (24h) → Final email with offer (48h). Data: cart_contents, cart_value, email, consent. Benchmark: open 50%, CTR 7%, CVR 8% (MailerLite 2025 + conversion benchmarks). Readiness depends on SFMC data extension setup.

Example 2 — Lead nurture B2B (AJO): User: "Design a lead nurture for AcmeAuto dealer inquiries on AJO. Leads come from website form." → 5-step journey: Welcome + resource → Case study (5d) → Engagement decision → Demo CTA or alternative content → Route to sales or long-term nurture. Data: lead_score, inquiry_type, dealer_region. AJO-specific: streaming segment for real-time entry, consent policy mandatory.

Example 3 — Winback (SFMC): User: "Customers who haven't purchased in 90+ days. SFMC. Telecom industry." → 4-step journey with benchmark adjustment: telecom open rates 38-42% (vs 43% global). Entry: inactivity trigger from DE filter. Channels: email → SMS → email with offer → exit to suppression or reactivation.

Skill Chaining

Direction	Skill	Connection
Upstream	audience-segmentation-brief	Provides segment definitions and data requirements
Upstream	media-routing-planner	Identifies CRM as a channel → triggers journey design
Upstream	scope-audit	Validates CRM journey is in SOW
Downstream	campaign-launch-qa	Journey spec → platform QA checklist (Step 6.1 tech sweep)
Downstream	weekly-control-tower	CRM KPIs feed weekly monitoring
Downstream	measurement-incrementality	Journey → incrementality test design
Downstream	executive-growth-memo	Journey results → leadership summary