map Service Blueprint settings_suggest DesignOps hub Cross-Team Governance

CX Blueprint — Foleon / AEM Content Process

Designing a 10-phase service blueprint that standardizes how Marsh delivers digital content across three technical scenarios — standalone Foleon, AEM-integrated, and AEM-embedded. Co-designed with Claude as AI collaborator to structure governance, QA protocols, and cross-team workflows.

10
Phases governing the full content lifecycle — from scenario determination through post-launch retrospective. 3 scenarios, 3 governance gates, 6 service layers, and a P1/P2/P3 QA severity model built into every delivery.
CX Service Blueprint — 10-phase Foleon/AEM content delivery framework

Client

Marsh — Insurance (Enterprise)

My Role

Lead CX Designer / Process Architect

AI Collaborator

Claude (Anthropic) — blueprint structuring & documentation

Platforms

Foleon · AEM (Adobe Experience Manager)

Tools

Wrike · Adobe Analytics · Figma · BrowserStack · Osano

Deliverables

Service Blueprint · 3-Scenario Framework · QA Protocol · Governance Gates · Wrike Templates

No Shared Process for Enterprise Content Delivery

Marsh's digital content team used Foleon and AEM to produce and publish content — but depending on the project, the technical architecture was completely different. Standalone Foleon assets, Foleon pages linked within AEM journeys, and Foleon content embedded as iframes in AEM all required different setups, different QA, different SEO approaches, and different rollback plans. Without a shared process, every project started from scratch.

Teams operated in silos. CX, SEO, Analytics, Platform, and QA each had their own timelines with no formal handoff points. Late stakeholder feedback regularly derailed launches. There was no severity classification for QA findings — a typo and a broken embed were treated the same way.

device_hubThree incompatible scenarios

Standalone Foleon, AEM-linked, and AEM-embedded content required entirely different technical setups — but were treated as the same type of project.

do_not_disturb_onNo governance gates

Stakeholders could request structural changes mid-production. Without formal approval checkpoints, late feedback continuously derailed timelines and forced costly rework.

groups_3Siloed teams

CX, SEO, Analytics, Platform, and QA each worked sequentially rather than in parallel. No shared language, no unified timeline, no handoff standards.

bug_reportUndefined QA severity

Without a severity model, every QA finding was treated as a blocker. This created ambiguity about what could delay a launch and what could be addressed post-live.

adjustMy Brief

Design a service blueprint that governs all three content delivery scenarios — creating a shared process language between CX, SEO, Analytics, Platform, QA, Designer, and stakeholders. The blueprint needed to define ownership, parallel workstreams, governance gates, and escalation paths from scenario selection through post-launch retrospective.

Strategic Choices That Shaped the Blueprint

Decision What I Chose Why Tradeoff
Blueprint format NN/g service blueprint (phases × layers) Different roles see the same process from their own layer — CX sees the journey, platform sees the tech, QA sees their stage. One document, multiple audiences. More complex to author, but far more durable than a simple checklist
Scenario architecture 3 distinct scenarios (A, B, C) with shared core phases A single "content delivery process" doesn't exist — the technical reality differs too much. Naming scenarios creates shared vocabulary without over-engineering. More documentation per scenario, but eliminates ambiguity on every project
Team workstreams Parallel tracks in Phase 3 (SEO + Analytics + Platform) Sequential handoffs added 1–2 weeks to every project. Running SEO, Analytics, and Platform planning simultaneously cuts calendar time without cutting depth. Requires stronger coordination, but CX brief in Phase 2 serves as the synchronization point
Governance model 3 formal gates: Gate 1, Gate 2, Go/No-Go Gates create shared commitment moments. Once a stakeholder approves Gate 1, structural changes require a formal change request — protecting the team from scope creep. Stakeholders must engage earlier, but late-stage surprises drop dramatically
QA severity model P1 Blocker / P2 Must Fix / P3 Improvement Without severity classification, any finding could theoretically block a launch. The model gives QA a shared language with IT/PM and stakeholders about what actually gates publication. Requires upfront alignment on definitions, but removes ambiguity in every QA conversation
AI collaboration Claude as co-designer for blueprint structuring Designing a process this complex involves generating, iterating, and reorganizing large amounts of structured content. Claude served as a thinking partner — drafting layer content, stress-testing scenarios, and formatting documentation. AI drafts need human review for accuracy, but the collaboration compressed weeks of process writing into days

10 Phases × 6 Service Layers

The blueprint uses the NN/g service blueprint model: read each column as a phase in the timeline, read each row as a layer of the service. Every cell is owned by a specific team and produces a specific deliverable or action.

10-phase service blueprint grid — phases vs service layers

Read each column as a phase · Read each row as a service layer · Gates mark formal approval moments

folder_open Evidence / Deliverables

What gets produced at each phase

From scenario decision doc and Wrike project brief through QA report, SEO validation, consolidated feedback, published URL, and final version log.

person Stakeholder Journey

What stakeholders experience

Scoping call → request submission → Gate 1 approval → progress updates → preview review → Gate 2 sign-off → 48hr final review window → written Wrike approval → performance report + retrospective.

support_agent Frontstage (Employee Actions)

What the team does that stakeholders see

IT/PM facilitates scoping; CX presents recommendations; Designer shares preview; QA communicates findings; SEO/Analytics report validation results; IT/PM confirms go-live; Analytics shares performance dashboard.

computer Technology Stack

Tools per phase

Wrike (project mgmt + automation) · Figma (journey mapping, A11y scanning) · Foleon (builder, publishing, SEO) · Adobe Analytics + Looker · AEM (scenarios B & C) · BrowserStack (device testing) · Claude (AI assist) · Osano (cookie consent) · info.marsh.com (publishing).

settings Backstage (Internal Actions)

What happens behind the scenes

Scenario evaluation → Wrike project setup → CX brief distribution → parallel SEO/Analytics/Platform planning → WCAG AA design + Osano/LCPA config → P1/P2/P3 QA → metadata validation → feedback implementation → pre-publication checklist → 72-hour monitoring.

library_books Support Processes

Standards that enable consistency

Scenario decision framework · Wrike automation templates · Design system + WCAG reference · SEO guidelines + analytics playbook · Foleon templates + naming conventions · QA checklist + device matrix · Feedback protocol + change request form · Retrospective template + version log.

0

Scenario Determination

IT/PM + CX evaluate the request using the decision framework. 30-min scoping call if unclear. Output: Scenario A, B, or C confirmed.

1 day

1

Intake & Project Setup

IT/PM creates Wrike project with 10 core tickets + scenario-specific additions. Brief submitted with objectives, audience, source materials, launch date, and distribution strategy.

1–2 days

2

CX Strategy Review

CX evaluates IA, content hierarchy, user journey, and mobile-first approach. CX brief distributed to SEO, Analytics, and Designer. → Gate 1

2–3 days

3

Parallel Planning

Three tracks run simultaneously: SEO (titles, metadata, canonical strategy) · Analytics (event tracking, dashboard setup) · Platform (hosting, AEM config, CSP/security).

2–3 days (parallel)

4

Design, Preview & Full Build

Designer creates WCAG AA-compliant Foleon asset, Osano/LCPA config, naming conventions. Preview review. → Gate 2. Full build incorporating SEO tags, analytics, and all metadata.

5–10 days

5

QA — Internal Review

Full checklist: accessibility, branding, device matrix (1920/1440/768/375px), content accuracy, links, performance. P1 blockers loop back to designer. Parallel: SEO + Analytics validation.

2–4 days

6

SEO & Analytics Validation

Parallel validation of metadata, alt text, indexing settings (SEO) and event triggers, scroll tracking, data flow (Analytics). Scenario C embed tracking is a critical test point.

1–2 days (parallel)

7

Stakeholder Review

48-hr review window. One designated feedback contact. Max 2 rounds. Feedback categorized as content correction, design adjustment, or new requirement (triggers change request).

2–5 days

8–9

Approval & Go-Live

Final QA confirms no P1 items. Written Wrike approval captured. Pre-publication checklist completed per scenario. → Go/No-Go. Publication with scenario-specific rollback plan on standby.

1–2 days

10

Monitoring & Retrospective

72-hour monitoring of traffic, events, performance, SEO indexing, and embed stability (Scenario C). 30-min retrospective within one week. Version log maintained for recurring assets.

72hrs + retro

One Blueprint, Three Delivery Architectures

The core 10-phase process applies to all projects. Scenarios define which technical path each project follows — determining hosting, SEO ownership, embed strategy, AEM involvement, and rollback complexity.

3 content delivery scenarios: Standalone Foleon, AEM Integration, AEM Embed
articleScenario A — Standalone Foleon

Foleon asset published directly on info.marsh.com. Full creative control. URL-based SEO owned entirely by Foleon.

  • Hosting: info.marsh.com
  • SEO: Foleon URL owns all signals
  • Rollback: Unpublish immediately (simple)
  • Osano + LCPA: Required on Foleon side

Complexity: Low

linkScenario B — AEM + Foleon Integration

Foleon content linked from AEM pages — both platforms form a connected user journey. Each platform owns its SEO. Design continuity across both is critical.

  • Journey: AEM → Foleon → AEM flow
  • SEO: Both URLs validated; canonical defined
  • Rollback: Take down both (sequence defined)
  • QA: Cross-platform re-test if either changes

Complexity: Medium

codeScenario C — Foleon Embedded in AEM

Foleon content embedded as an iframe within AEM pages. Most technically complex — CSP headers, cookie consent, embed scroll behavior, and analytics firing all require specific validation.

  • SEO: AEM page only (iframe not indexed)
  • Critical test: Embed tracking in AEM staging
  • Rollback: Replace embed with static placeholder
  • No scroll-within-scroll designs allowed

Complexity: High

lockThree Approval Moments That Protect the Team

⛔ Gate 1 — After Phase 2

Content Structure & Journey Approval

Stakeholder confirms: content structure, user journey direction, scenario approach. No design work begins without this approval. Post-gate changes require a formal change request.

⛔ Gate 2 — After Phase 4 Preview

Content Substance Approval

Stakeholder confirms: content completeness and accuracy, structure correctness, navigation functionality, visual direction. Post-gate changes require change request with timeline impact assessment.

⛔ Go/No-Go — After Phase 8

Pre-Publication Checklist Complete

All scenarios: Wrike closed, metadata verified, analytics live. Scenario-specific: hosting confirmed (A), both platforms ready (B), AEM staging validated + CSP confirmed (C).

From Ad-hoc to Governed Delivery

Delivery Process

Ad-hoc arrow_forward 10 phases

Structured

end-to-end blueprint

Scenarios

Undefined arrow_forward 3 scenarios

Classified

A · B · C with own protocols

QA Severity

All blockers arrow_forward P1/P2/P3

Model

shared severity language

Approval Gates

None arrow_forward 3 gates

Governed

G1 · G2 · Go/No-Go

map
Service BlueprintNN/g-style blueprint covering all 10 phases × 6 service layers — available as PDF and two interactive HTML versions
device_hub
3-Scenario Decision FrameworkEvaluation criteria for determining Scenario A, B, or C — including a 30-min scoping protocol for ambiguous cases
lock
Governance Gate ProtocolFormal approval checkpoints at Gate 1, Gate 2, and Go/No-Go with change request procedures and timeline impact assessment
checklist
QA Severity ChecklistP1/P2/P3 classification model with device matrix (1920/1440/768/375px), browser list, and scenario-specific test cases
task
Wrike Project Templates10 core tickets + scenario-specific additions for A, B, and C — with automation triggers and assigned owner placeholders
analytics
SEO & Analytics Validation ProcessParallel validation tracks with scenario-specific requirements — including embed tracking as critical test for Scenario C
A blueprint is a shared language, not a checklist. The most valuable part of the NN/g format isn't the phases — it's that each team can read the same document from their own row. CX reads the stakeholder journey. Platform reads the tech stack. QA reads the backstage. One document serves everyone without simplifying for anyone.
AI as a design thinking partner changes how process design works. Using Claude to structure, draft, and stress-test the blueprint wasn't about speed — it was about thinking out loud at scale. When you're designing a process with 10 phases × 6 layers × 3 scenarios, you need a collaborator who can hold all that complexity simultaneously and surface gaps you haven't seen yet.
Parallel tracks are a scheduling decision, not a process shortcut. Running SEO, Analytics, and Platform planning simultaneously in Phase 3 doesn't skip anything — it reorganizes the work onto a shared calendar. The CX brief becomes the synchronization artifact. This one change saved 1–2 weeks on every delivery without removing any step.
Gates protect the team, not just the stakeholders. Gate 1 and Gate 2 are usually framed as stakeholder checkpoints. But their real value is protecting the designer and QA team from late structural changes. Once a gate is signed, changes require a formal request — creating accountability on both sides of the relationship.
Naming scenarios creates operational clarity instantly. Before the blueprint, every project was "a Foleon project." After it, team members could say "this is a C" and immediately everyone knew: CSP headers, AEM staging, embed tracking critical, complex rollback. Shared vocabulary compresses every kickoff conversation.