How He Makes $60K/Month With AI Marketing (Claude Code)

Business outcome & commercial model

• Mateo Senzone runs a mostly automated one-person AI marketing agency.
• Revenue / pricing (reported):
  • $7k–$10k per client (average bracket)
  • ~$60k/month overall
• Capacity / time:
  • ~4 hours/week total across 6 SEO + Meta ads clients
  • Historically, backlink outreach alone required 5–8 hours/day (manual) before automation.

Core strategy: automate “full-funnel” marketing execution with AI agents

He positions his system as a “full stack marketer” approach:

• Top of funnel
  • Meta ads
  • Google ads
  • Cold outreach
  • Influencer outreach
• Retargeting / warm audience
  • Meta retargeting via pixel
• SEO growth
  • Backlinks/outreach to earn link placements
• Lifecycle
  • Email list / email marketing (mentioned as part of the stack)

Automation playbooks & frameworks (as described)

Agent-based workflow automation

He uses specialized agents for:

• SEO agent
  • Backlink outreach negotiation + sending outreach
• Meta ads agent
  • Campaign + creative + setup
• Google ads agent
  • Mentioned as part of the agent set

He grounds execution using an “AI agnostic brain” (Obsidian) to reduce failures and keep context consistent.

Competitive research loop (creative strategy learning)

• Use Meta Ads Library to inspect competitor creatives and copy.
• Run an AI workflow to:
  • download/transcribe competitor ads
  • store findings in Obsidian
  • derive strategy

Prompt engineering as a process improvement

• “Prompt Engineer” skill turns a rough request into structured prompts using:
  • the right context
  • client account data
  • KPI definitions
  • access paths

Retargeting as the low-budget demand capture tactic

• Install a Meta Pixel to build a warm audience.
• Run retargeting with small daily budgets.

KPIs & targets mentioned

• Lead / pipeline example (luxury real estate client):
  • ~35 leads from one AI-made ad
  • ~50% lead qualification rate (described as “really looking for a property”)
  • Closed deals: 3–4 people (for million-dollar properties)
  • Context note: a “4 million property” implies very high ticket value
• Retargeting budget guidance (general):
  • $3–$5/day is cited as enough to see results
• Operational KPI:
  • 4 hours/week serving 6 clients

Concrete examples & how the system works (execution details)

1) SEO: automated backlink outreach + negotiation

Belief: SEO rankings depend heavily on backlinks.

Manual process pain (historical):

• outreach
• price negotiation
• back-and-forth with site owners

AI-agent implementation:

• Maintains a list of 1,000+ sites and their site masters / intermediaries
• Agents:
  • email outreach
  • negotiate and “lowball” to get best pricing (stated candidly)
  • automatically continue/adjust negotiation if a site owner responds

2) Meta ads: creative generation + campaign setup automation

Main barrier (historically):

• building ads manually (Canva/Photoshop), plus time/cost overhead

AI creative engine:

• “Meta Ads Creative Engine” plus tools like:
  • Higsfield (paid)
  • free tooling including Remotion + FFmpeg (per the host)
• Inputs: client website + assets
• Output: hundreds of ad creatives (if desired)
• Workflow:
  • generate creatives quickly with prompts
  • feed creatives into Playwright-based automation (“Playriter”) to set up campaigns

Creative performance example:

• Luxury real estate ad with AI video-style animation of Tokyo, then appended logo/text
• Result (high-ticket context):
  • ~35 leads
  • 3–4 closed

3) Retargeting: Meta Pixel + warm audience ads

• Pixel purpose:
  • track visitors who don’t convert immediately
  • enable retargeting
• Positioning:
  • warm audience = prior visitors/actions
  • cold audience = first-time viewers
• Claim:
  • can run for $3–$5/day

4) “Command center”: always-on remote AI operations

Obsidian is used as a persistent knowledge store (“AI agnostic brain”):

• Sources sync into Obsidian via always-on “Daemons”:
  • Slack
  • WhatsApp
  • Discord
  • Gmail
  • (selected conversations/contacts)
• Benefit: avoids “vendor lock-in” risk and centralizes history for AI context.

Command center setup:

• Multiple “headless terminals” run continuously, labeled per client.
• When a message arrives:
  • an agent reads/relays the task
  • he chooses urgency
  • he authorizes action from phone
  • the agent completes browser-based tasks

Token/context issue mitigation:

• He emphasizes failures happen when AI lacks full context; Obsidian helps reduce that risk.

5) Browser automation for marketing and ops (Playwriter / Playriter)

He distinguishes between:

• Playwriter Chrome extension
  • works on the real signed-in browser
• Playwright MCP (“Playr Writer”/MCP variant)
  • can spawn its own browser
  • may cause login/sign-out issues

Use cases:

• Automate Meta ad creation steps:
  • click create
  • upload ads/assets
  • paste copy
  • run without manual keyboard work

Non-marketing examples (to show general capability):

• buy items on Amazon based on a photo request
• inquire with multiple lawyers’ websites to generate leads/referrals

Security guidance:

• Use a password and optionally restrict by IP.
• Warns that without authentication, someone could control the machine.

Actionable recommendations (implicit “how to start” path)

• If you’re new to marketing: study first, then use AI agents.
  • He recommends 1–3 months to reach “basic understanding,” depending on time investment.
  • Avoid jumping straight into cold outreach without fundamentals—generic outreach is often ineffective because many prospects have seen “people trying to sell SEO/meta ads.”
• Train agents only with real expertise content
  • Don’t “plug-and-play” an agent with zero marketing knowledge; accuracy depends on training data/inputs.
• Build infrastructure before scaling client work
  • Set up:
    • Obsidian knowledge base
    • daemons for syncing business communications
    • a command center to execute tasks remotely
  • Use “prompt engineering” to standardize prompts and reduce context errors.

Notes on marketing theory shifts (Meta specifics)

• “Creative is everything” on Meta:
  • targeting via interests is largely gone / less important now
  • Meta ad delivery is influenced more by what the ad says and shows
  • Suggested implication: prioritize compelling offers + clear creatives over complex demographic targeting.

Presenters / sources

• Presenter: Mateo Senzone (also referred to as “Murasaki Mateo” in the outro)
• Video host/organizer: Not explicitly named in the subtitles (speaks as “Sandy” in one exchange)