Marketing Operations · AI-Augmented
Marketing
at Machine Speed
How a three-person marketing team and Claude ship an entire campaign pipeline simultaneously without adding headcount.
The Collaboration Model
A New Kind of Creative Partnership
A loop where human strategy and AI execution multiply each other.
Write
CMO writes the campaign white paper by hand — friction drives original thinking
Develop
Claude and CMO refine the white paper draft together into the final asset
Expand
Every other campaign asset generated from that finished white paper
Deploy
20+ assets published across web, email, and partner channels
7+
Active campaigns in flight
5+
Asset types per campaign
3
People in the entire team
247
MQLs delivered Q1 2026
Active Campaigns
Multiple Campaigns. Simultaneously.
Traditional marketing runs one campaign at a time. This approach keeps seven active simultaneously, each at a different lifecycle stage.
Live
NVIDIA VergeOS AI Workstation
Press ReleaseLanding PageBattlecardWebinarEmailBlog
Launching
Veeam & VergeOS oVirt Migration
Landing PageSolution BriefPartner PageEmail
Live
Flash & Memory Supercycle
White PaperOn-Demand WebinarBlog Series
Building
VergeOS Performance vs. vSAN
Benchmark ReportCompetitive PageDemo Brief
Launching
Storware Integration Webinar
Reg PagePartner EmailSlide Brief
Live
Dell Private Cloud vs. VergeOS
Competitive BriefTCO CalculatorBlog
The Creative Origin
Every Campaign Starts With a Blank Page
Not a prompt. Not a brief template. A blank page and a pen. The CMO writes every campaign white paper before Claude is involved at all.
The Founding Principle
Creativity is benefited by friction.The white paper is written first.
Every campaign begins the same way: the CMO sits down and writes. Typed or hand-written. What matters is that the thinking happens without AI assistance. Wrestling with a blank page produces original positioning and honest differentiation that conversation alone cannot surface.
Why the friction matters
When you write without AI, you hit the hard parts: claims you can’t fully support, positioning that doesn’t quite land, technical nuance not yet resolved. Those friction points reveal exactly where the campaign needs to go deeper. That raw material is more valuable than a smooth AI-generated first draft.
Every asset below flows from that one white paper:
Press Release
Campaign Blogs
Landing Page
Sales Battlecard
Sales Guide
Webinar Outline
Email Sequences
StorageSwiss Articles
VMblog Posts
External Placements
ChalkTalk Video Brief
Case Study Frame
Post-Webinar Blog
6-Week Follow-on
How the white paper becomes a campaign
1
Human only
CMO writes the rough draft
Typed or hand-written. No prompts, no AI. The goal is honest thinking: technical arguments, competitive framing, use cases, positioning — including the gaps and uncertainties.
2
Human + Claude
Collaborate to finalize the white paper
The rough draft enters Claude. Structure tightened, claims stress-tested, brand voice applied, argument made airtight. The most intellectually intensive part of the process.
3
Claude + Human review
Generate all campaign assets from it
The white paper becomes the campaign source of truth. Every other asset adapts its established positioning to each format and audience. Nothing starts from scratch again.
4
Human
Final approval on every asset
Every generated asset reviewed by the CMO before publish. Technical accuracy, competitive positioning, and relationship judgment stay human.
Quarterly Asset Output
The Asset Factory
Quarterly numbers across all active campaigns simultaneously — not per campaign. Claude generates the first draft of every asset, tuned to the same voice and positioning.
12+
Blog Posts
Under 1,000 words
Technical, opinionated
SEO-optimized
Technical, opinionated
SEO-optimized
6+
White Papers
3,000–4,000 words
Deep-dive technical
Gated PDF assets
Deep-dive technical
Gated PDF assets
8+
Landing Pages
Full HTML + CSS
Campaign reg pages
Webinar & gated
Campaign reg pages
Webinar & gated
5+
Battlecards
Competitive intel
Sales enablement
Partner portal
Sales enablement
Partner portal
20+
Emails
Nurture sequences
Launch announcements
Partner outreach
Launch announcements
Partner outreach
4+
Press Releases
750-word format
VMblog-ready
Distribution copy
VMblog-ready
Distribution copy
7+
Webinar Decks
Briefs & outlines
Speaker notes
Follow-up content
Speaker notes
Follow-up content
2+
Calculators
HTML tools
TCO & infra sizing
Interactive plugins
TCO & infra sizing
Interactive plugins
The Integration Layer
MCP: Claude Inside Your Apps
Model Context Protocol turns Claude from a chatbot into a connected agent — reading, writing, and acting across every local and remote tool without manual handoffs.
Model Context Protocol
What MCP Actually Is
The open standard that lets Claude connect to local Mac apps and remote services through live, bidirectional data access running on your machine.
1
Local-first, not cloud-dependent
MCP servers run on your Mac. Things, DEVONthink, Bear, Drafts, Fantastical, and Apple Mail connect via local MCP servers. Claude talks to them directly, keeping data on-device with no third-party relay.
2
Read and write, not just read
Claude creates tasks in Things, writes records to DEVONthink, pushes pages to WordPress, and drafts Slack messages. Every action that used to require a tab switch is now a single prompt.
3
Composable across tools in one session
Claude can read a task from Things, pull context from DEVONthink, generate content, push to WordPress, create a follow-up task, and draft the Slack message — all in one conversation.
4
Each server adds a capability layer
Installing an MCP server gives Claude a full set of operations for that app. Every server you add expands what Claude can autonomously do.
One prompt. Six tools. Zero copy-paste.
Things 3
Local · Mac MCP
MCP
Reads today’s campaign tasks and deadlines
DEVONthink
Local · Mac MCP
MCP
Pulls yesterday’s dashboard and prompt library
WordPress
Remote · Claudeus MCP
MCP
Pushes approved HTML to live pages
Apple Mail
Local · Mac MCP
MCP
Reads and triages inbox across four accounts
Slack
Cloud · API MCP
MCP
Drafts team announcements for review
Safari
Local · Browser MCP
MCP
Live research, Salesforce MQL reads, screenshots
Task & Knowledge
Things 3
read · write · update
DEVONthink
read · create · archive
Bear Notes
read · create · tag
Drafts
create · run actions
NotePlan
read · create notes
Agenda
read · create
Communication
Slack
read · draft · send
Apple Mail (read)
read · search · link
Apple Mail (write)
compose · archive
Mac Messages
read · search
Fantastical
read · create events
Day One
read · journal
Web & Publishing
WordPress (Claudeus)
pages · posts · media
WordPress.com
StorageSwiss blog
Safari MCP
browse · read · click
Web Search
real-time research
Code Execution
build · test · deploy
Why Local Apps Change Everything
Most AI workflows connect to cloud services. This one connects to local Mac applications — the apps where the actual work lives. Claude has access to the complete working context: every task in Things, every archived document in DEVONthink, every drafted note in Bear, every email thread in Apple Mail. Not a cloud summary — live, native data, on-device. The result: Claude knows what you worked on yesterday, what’s due next week, and what you half-wrote in Drafts last Tuesday. Full situational awareness, every session.
The Tech Stack
Hub-and-Spoke, Not Copy-Paste
Claude sits at the center — reading from local apps on the left, publishing to web and comms tools on the right, all inside one daily conversation.
✅
Things 3
Task & campaign tracking
🗄️
DEVONthink
Knowledge base & archive
✏️
Bear + Drafts
Writing staging & review
🐻
NotePlan
Calendar-linked notes
📅
Fantastical
Event & deadline mgmt
CLAUDE
+ MCP
+ MCP
THE HUB
🌐
WordPress
Direct page publishing
💬
Slack
Team communication drafts
📬
Apple Mail
Email read, compose, archive
🦁
Safari
Web research & Salesforce
📅
Day One
Daily ops journaling
Daily Operations
A Day in the System
The same loop runs every workday. Structured, repeatable, and fast.
01
Human + Claude
Daily Dashboard Build
Each morning: Claude walks active Things projects, reads the prior day’s DEVONthink Final record, checks deadlines, surfaces blockers, and builds a prioritized view — all from a single stored prompt.
02
Human
CMO Writes the White Paper First
Every campaign starts with a white paper written by the CMO — typed or hand-written. No prompts, no AI. Just a blank page and original thinking. The friction is intentional: wrestling with an idea produces sharper positioning than prompting ever could.
03
Human + Claude
Collaborate to Finalize the White Paper
The rough draft enters Claude as a working document. Structure, technical accuracy, brand voice, argument — all refined together. The result carries the CMO’s original thinking honed to publication standard.
04
Claude
Generate All Campaign Assets From the White Paper
The white paper is the source of truth. Blogs, press release, battlecard, landing page, email sequences, sales guides, webinar outlines, external articles — all generated from it. No asset starts from scratch.
05
Claude
Publish & Park
Once approved: push HTML to WordPress via Claudeus MCP, archive to DEVONthink, create follow-on tasks in Things, draft the team Slack announcement. Session ends with a structured parking prompt — everything logged, nothing lost.
Beyond Content
Software Built, Not Just Content Written
Claude doesn’t just generate words. The team has shipped functional marketing software — tools that run on the live site and serve visitors every day.
Full campaign planning calendar in WordPress. Every campaign, deadline, launch date, and owner in one view. Built from scratch — zero third-party dependencies.
Drag-and-drop campaign scheduling
Color-coded by status and owner
Synced with task tracker
Month, quarter, and list views
Custom project and task management inside WordPress. Tracks assets from brief through publish with a full audit trail and remote API for cross-system integration.
Remote POST API for external integration
Assignee tracking by team member
Campaign-level rollup views
Brief → Draft → Review → Live pipeline
Monthly budget tracking in DEVONthink. Pulls actuals, calculates variance, generates a formatted spend report on demand. One prompt builds the full view.
Category-level budget vs. actual
Variance alerts and trend view
Prompt-generated HTML report
Stored prompt triggers full rebuild
Interactive ROI calculator on verge.io. Prospects model savings from running VergeOS on their own hardware vs. a hyperscaler. Built and deployed through the Claude workflow.
Dynamic sliders with live recalculation
Side-by-side cost comparison vs. cloud
Exportable summary for sales
Mobile-responsive, VergeIO design system
The Prompt Repository
Memory as Infrastructure
Storing proven prompts in DEVONthink as reusable assets — a living library that compounds over time.
🗄️
DEVONthink · AI Prompts Library
Always current
Daily Dashboard Build v2
Daily Ops
Campaign Launch Checklist
Campaign
Inbox Triage Directive
Email
Battlecard Generator
Competitive
Things Triage v2
Task Mgmt
Landing Page Build
Web
White Paper Framework
Content
Daily Shutdown Routine
Daily Ops
Prompts are assets, not instructions
A well-written prompt in DEVONthink is as valuable as any other marketing asset. Version it, refine it, treat it like production code — it drives production output.
One prompt per repeatable workflow
Every repeating operation has its own stored prompt. Zero re-explanation. Copy, paste, run. Consistent input produces consistent output.
Brand rules live in the prompt
Style guidelines, banned words, sentence limits — embedded in every content prompt. Claude never needs re-briefing. The prompt carries the knowledge.
Prompts improve over time
When a prompt produces output that needs fixing, update the prompt — not just the output. Each refinement improves every future session that uses it.
Extended Thinking
Watch Claude Think — and Use What You See
Extended thinking exposes the reasoning chain — often more valuable than the output itself.
What Extended Thinking Actually Does
Claude reasons through the problem before answering — working through assumptions, weighing tradeoffs, stress-testing logic. You see that chain stream in real time. For complex decisions, watching the reasoning reveals angles you wouldn’t have considered yourself.
Catch Bad Assumptions Early
The thinking stream reveals when Claude is working from a flawed premise. Interrupt and correct before it builds an entire asset on the wrong foundation.
Steal the Reasoning, Not Just the Output
The thinking block surfaces competitive angles and positioning logic that never makes it into the final draft. Those raw thoughts are a strategy document.
Use It for Complex Decisions, Not Simple Tasks
Extended thinking adds time. Reserve it for strategy and competitive analysis. Standard mode is faster for routine content generation.
Workflow in Depth
What Each Party Actually Does
Human judgment drives decisions. Claude drives execution. The system ensures nothing falls through the cracks.
The Human Layer
Writes the White Paper First
Every campaign starts with the CMO writing — no AI. Friction is deliberate: a blank page produces sharper thinking than prompting.
Market Intuition
Reading competitive signals, sensing market timing, knowing which angle lands with which buyers — before a word goes into Claude.
Technical Accuracy
Every technical claim reviewed by a former analyst before anything goes live. Claude generates; the human verifies.
Relationship Calls
Which partner gets which content, how to handle a sensitive reference, when to push and when to hold.
Prompt Engineering
Writing and refining the prompt library. Its quality directly determines output quality.
The Claude Layer
White Paper Collaboration
Works with the CMO to refine rough drafts into publication-ready white papers that anchor each campaign.
Asset Generation
Adapts the white paper into every campaign format — blogs, battlecards, emails, landing pages — in brand voice.
Tool Orchestration
Reading from Things, writing to WordPress, archiving to DEVONthink, drafting in Slack — in a single session.
Code & App Development
Building WordPress plugins, interactive calculators, and custom dashboard systems from functional specs.
Session Memory
Holding full campaign context — brand rules, prior decisions, constraints — without re-briefing.
The System Layer
MCP Tool Connections
Model Context Protocol bridges Claude to every app — eliminating manual handoffs.
DEVONthink as Brain
Every asset, prompt, and dashboard archived. Grows with every session and feeds every future one.
Things as Command Center
Campaign tasks and deadlines in Things. Claude walks the projects each morning for the day’s view.
Parking Protocol
Every session ends structured: tasks parked, assets archived, drafts staged. Nothing lost between sessions.
One Conversation Per Day
All work in a single daily conversation — maintaining context integrity throughout.
The Insight
Why This Works
Friction Creates Originality
Every campaign starts with the CMO writing alone — no AI, no templates. The struggle of a blank page produces genuine insight that polished prompting never surfaces. Claude refines that thinking; it doesn’t replace it.
One Source, Twenty Assets
The finished white paper is the campaign’s single source of truth. Every blog, battlecard, email, and landing page is an adaptation of it. Positioning stays consistent. Claims stay verified. Voice stays right.
Tools Eliminate Handoffs
MCP connections mean Claude reads from task managers, writes to WordPress, and archives to the knowledge base directly. The copy-paste bottleneck — the biggest time sink in manual workflows — is gone entirely.
Practitioner Tips
What Makes This Actually Work
Practices refined over months of daily use that separate productive AI collaboration from frustrating trial and error.
01
Write the white paper without AI first
Resist the urge to start with a prompt. Write the campaign argument by hand. The friction produces positioning and insight that conversation cannot generate. That rough draft is more valuable than a polished AI-written one.
02
Start every session with full context
Open each session with the daily dashboard prompt — it pulls current task status, active campaigns, and prior decisions into context. The first 60 seconds determines the quality of the next two hours.
03
Build your style rules once, embed them everywhere
Voice guidelines, banned words, sentence length limits belong in the prompt. Document them in a
Brand Voice block in DEVONthink and paste them into every content prompt.04
Review drafts in chat before archiving anywhere
Show the draft inline, collaborate in conversation, finalize — then push to WordPress and archive to DEVONthink. Never publish from a first-pass.
05
Use extended thinking for strategy, not speed
Turn on extended thinking for campaign positioning and competitive response. Standard mode is faster and sufficient for routine content generation.
06
Name your snapshots and park at end of day
Intraday saves go to DEVONthink as
Dashboard MM-DD at HH:am/pm. End of day becomes MM-DD-YYYY-Final. Tomorrow’s session pulls from that Final record.07
Treat the prompt library like production code
When a prompt produces output that needs fixing, update the prompt — not just the output. Every future session using it gets the benefit of every past refinement.
08
Three people, no excuses
The team is three people. Not adding headcount. Every new campaign type gets a reusable prompt template. Every new workflow gets an MCP integration. The leverage compounds. The team scales without scaling.
Sample Campaign
What One Campaign Actually Looks Like
Refurbished SSD: Beat the Supercycle, Safely — April through June 2026. 74 tasks across 20 asset types, managed inside the Claude workflow.
Campaign at a Glance
74
Total tasks
20
Asset types
10
Weeks active
5
Publications
1
Live webinar
3
People
How Claude Compresses This Timeline
This campaign would traditionally require 6–8 weeks of production with a larger team. With Claude, first drafts of every written asset are ready within the first two weeks. The team shifts from production to review, refinement, and relationship management.
Asset / Phase
Apr W1
Apr 1–7
Apr 1–7
Apr W2
Apr 8–14
Apr 8–14
Apr W3
Apr 15–21
Apr 15–21
Apr W4
Apr 22–28
Apr 22–28
May W1–2
Apr 29–May 12
Apr 29–May 12
May W3–4
May 13–31
May 13–31
June
Jun 1–7+
Jun 1–7+
Campaign BriefStrategy foundation
Press ReleaseLaunch announcement
White Paper3,000–4,000 words
Landing PageCampaign hub
Sales EnablementGuide, battlecard, slides
Blog Wave 1VergeIO, StorageSwiss, VMblog
Case StudyCustomer interview + draft
Blog Wave 2 + ExternalGhost, B&F, DCIG, VMblog
ChalkTalk VideoOutline → produce → post
WebinarLive Apr 23 · Reg page Apr 5
Post-Webinar VideoClip from live recording
6-Week Follow-onInsight blog + email
Strategy
PR / Webinar
Blog Placements
Web / Landing Page
Sales Enablement
Video
Case Study
Long-tail