Dejon Email Workflow Analysis
Comprehensive workflow documentation and AI automation roadmap
The email campaign workflow spans 14 distinct applications across 10 categories, with 3 overlapping email marketing platforms creating significant redundancy.
The email campaign lifecycle flows through 6 major phases with 25+ application transitions, each requiring manual context switching and often re-authentication.
3 Email Platforms, 1 Job
SFMC, Mailchimp, and Klaviyo all handle overlapping email campaign functions. Campaign configuration is performed independently in each, creating 5-10 minutes of redundant work per campaign.
Manual Asset Shuttle
Images flow from Google Drive / local file system through manual download-upload cycles into SFMC and Mailchimp. No centralized DAM or automated pipeline connects storage to builders.
23 friction points identified across 9 categories. The top 5 account for an estimated 60-90 minutes of lost productivity per campaign session.
Tool Fragmentation + Platform Duplication
14 tools with 3 overlapping email platforms. 25+ context switches per session. 15-35 min lost per campaign.
Manual Newsletter Content Aggregation
Staff profiles, events, articles gathered from multiple sources by hand. 15-30 min per newsletter.
Disconnected QA Processes
Email on Acid as a separate auth-gated tool + Miro for visual comparison. 15-25 min per campaign.
Fragmented Asset Management
Assets scattered across local file system, Google Drive, and Workfront. 5-15 min searching per campaign.
Status/Feedback Outside Production
Miro stamps and annotations for campaign status and design feedback creates sync risk. 10-20 min cross-referencing.
View full friction catalog (23 items)
| ID | Category | Description | Severity | Time Cost | Automation |
|---|---|---|---|---|---|
| F-1.1 | Context Switching | 14 apps in single workflow, 25+ switches | High | 12-25 min/session | High |
| F-1.2 | Context Switching | 15-20+ browser tabs simultaneously | Med | 1-2 min/session | Med |
| F-1.3 | Context Switching | Manual multi-window arrangement | Low | 1 min/occurrence | Med |
| F-2.1 | Duplicate Platforms | 3 email marketing platforms (SFMC, Mailchimp, Klaviyo) | High | 5-10 min/campaign | High |
| F-2.2 | Duplicate Platforms | Assets in 3 disconnected locations | High | 2-5 min/retrieval | High |
| F-3.1 | Authentication | Email on Acid repeated OTP | Med | 2-4 min/session | High |
| F-3.2 | Authentication | Salesforce session timeout | Low | 30-60 sec | Med |
| F-4.1 | Manual Data Transfer | Newsletter content aggregation from multiple sources | High | 15-30 min/newsletter | High |
| F-4.2 | Manual Data Transfer | Complex multi-step SFMC export | Med | 90 sec/export | High |
| F-4.3 | Manual Data Transfer | Manual asset download/upload between systems | Med | 1-3 min/asset | High |
| F-4.4 | Manual Data Transfer | Placeholder text manual replacement | Med | 5-10 min/newsletter | High |
| F-5.1 | Manual QA | Separate tool for email QA testing | High | 3-5 min/cycle | High |
| F-5.2 | Manual QA | Manual inbox rendering verification | Med | 2-3 min/send | High |
| F-5.3 | Manual QA | Visual QA via screenshots in Miro | High | 10-20 min/campaign | High |
| F-6.1 | Status Tracking | Status via Miro stamps instead of system fields | High | Ongoing risk | High |
| F-6.2 | Status Tracking | Feedback in Miro instead of production tools | Med | 5-10 min/review | High |
| F-7.1 | Performance | "Site can't be reached" error during workflow | Med | Variable | Low |
| F-7.2 | Performance | Loading delays across multiple tools | Low | 2-5 min/session | Low |
| F-7.3 | Performance | Cookie consent popup interruption | Low | 5-10 sec | Low |
| F-7.4 | Performance | Email builder frustration/difficulty | Low | Indeterminate | Med |
| F-8.1 | HTML Editing | Raw HTML editing for template customization | High | 5-15 min/session | Med |
| F-8.2 | HTML Editing | Deep folder navigation for asset insertion | Med | 1-2 min/insertion | High |
| F-9.1 | Clutter | In-app advertising in Klaviyo | Low | Minor | Low |
18 automation opportunities across 3 implementation horizons. Combined potential: 15.5-41.5 hours/month recovered for strategic and creative work.
HTML Email Code Generation
Claude generates HTML/CSS changes from natural language descriptions. Eliminates manual code editing in the WYSIWYG's code view.
SFMC Export Automation
Computer use or API bypasses the complex multi-step save/export dialog. Single command replaces 90-second manual process.
Session & Auth Management
API-based access eliminates OTP authentication. Session keepalive prevents Salesforce timeouts.
Pre-Send Checklist Agent
Claude validates subject line, preheader, from address, links, and images before every send. Catches errors before QA.
Template Variant Generation
Given a base template and regional variations, Claude generates all variant HTML files automatically.
Workfront Auto-Update
Agent detects campaign milestones and updates Workfront task statuses automatically.
Content Proofreading
Claude reviews email copy for grammar, brand voice, compliance, and accessibility before sending.
Accessibility Remediation
Claude auto-generates alt text, fixes contrast issues, and adds ARIA labels to email HTML.
Centralized Asset Management
MCP servers index Google Drive, local file system, and Workfront. Natural language search: "find the Colorado hero image." Auto-uploads to destination platform.
Automated Email QA Pipeline
Email on Acid API eliminates browser auth. Claude parses results, suggests fixes, auto-remediates routine issues, retests.
Variant Visual Comparison
Computer use captures screenshots of all variants. Claude Vision compares them and generates diff reports. Replaces Miro screenshot workflow.
Campaign Status Sync
Agent polls SFMC, Mailchimp, Klaviyo for status changes. Auto-updates Miro stamps and posts Slack notifications.
Miro Board Auto-Generation
Agent creates campaign review boards programmatically: screenshots, variant grids, status stamps, and annotation templates.
Brief-to-Template Kickoff
Claude reads the campaign brief, selects the best template, pre-populates content, and sets up the initial campaign in SFMC.
AI Newsletter Content Aggregation
Custom agent monitors content sources, extracts structured items, generates newsletter copy in brand voice, and pre-populates templates. Dee reviews and approves.
Cross-Platform Campaign Sync
Configure campaign once in SFMC. Agent propagates to Mailchimp and Klaviyo via MCP servers. Field mapping, audience validation, send-time conflict detection.
Unified Command Interface
Single chat-based interface orchestrates all 14 tools. "Show March P&C status across all platforms" queries SFMC, Mailchimp, Klaviyo, and Miro simultaneously.
Performance Learning Loop
Agent analyzes send results, identifies winning patterns, and recommends optimizations for future campaigns. Continuous improvement feedback.
Use the category tabs above to filter, or expand the sections below.
Quick Wins (8 opportunities)
Medium-Term (6 opportunities)
Transformative (4 opportunities)
Phased implementation ordered by impact-to-effort ratio, building capabilities incrementally.
Phase 1: Foundation & Quick Wins
Deploy Claude API for HTML code generation (A-06), template variant generation (A-13), content proofreading (A-15), and pre-send checklist (A-11). Set up API-based access to Email on Acid to eliminate OTP auth (A-09). Automate SFMC export workflow (A-07).
Phase 2: Integrated QA & Asset Pipeline
Build automated QA pipeline connecting SFMC to Email on Acid API (A-04). Deploy centralized asset management with MCP servers for Google Drive, local file system, and Workfront (A-03). Implement automated variant visual comparison (A-05).
Phase 3: Cross-Platform Orchestration
Build MCP servers for SFMC, Mailchimp, and Klaviyo APIs. Deploy cross-platform campaign sync agent (A-02). Implement campaign status synchronization (A-08). Auto-generate Miro review boards (A-12). Connect brief-to-template kickoff (A-16).
Phase 4: Unified Intelligence
Deploy AI-powered newsletter content aggregation (A-01). Build the unified command interface that orchestrates all tools from a single chat (A-10). Implement the performance learning loop for continuous optimization (A-18).
The Manual Marathon
- 14 tools with no integration
- 25+ manual context switches per session
- 3 email platforms doing the same job
- Manual HTML editing for customizations
- Separate auth-gated QA tool
- Screenshot-based visual review in Miro
- Status tracked with manual stamps
- 60-90 min lost per campaign to friction
AI-Orchestrated Workflow
- Single command interface for all tools
- Automated cross-platform campaign sync
- Claude-powered HTML generation
- API-based QA with auto-remediation
- Vision-based variant comparison
- Real-time status synchronization
- AI newsletter content assembly
- 2-5 working days/month recovered