Team Briefing

The Age of
Agentic AI

What it means for 11 Agency, how to use it,and where we're going.

Raymond · March 2026
Scroll ↓
The Shift

AI went from answering questions
to doing work.

This isn't ChatGPT. Agentic AI connects to your tools, remembers context, executes multi-step tasks, and gets better every time you use it.

Before — Chatbot Era
  • One prompt, one answer
  • Copy-paste into your tools
  • No memory between sessions
  • Generic output, needs heavy editing
  • Can't take actions or connect to anything
Now — Agentic Era
  • Multi-step execution across tools
  • Publishes directly to Klaviyo, Typefully, etc.
  • Persistent memory — learns your preferences
  • Quality improves with every correction
  • Reads files, writes code, calls APIs, manages data
What I Built

A full content system
powered by AI agents.

In less than a week, I built an orchestrated pipeline that generates, scores, and publishes content across LinkedIn, X, and our newsletter.

6,800
Lines of config
45
Files
10
AI agents
3
Platforms
Weekly Pipeline
1
Analyst
Pulls meetings, news, competitor intel
2
Calendar Strategist
Plans the week's content calendar
3
Platform Strategists
Write LinkedIn, X, and newsletter drafts
4
Chief Strategists
Score against rubrics, enforce voice
5
Publishing
Push to Typefully + Klaviyo at scheduled times
6
Engagement
Comment on targets, manage replies
7
LinkedIn posts/week
14
X posts/week
2
Newsletters/week
Real Output

Everything you've seen this week
was built by this system.

The LinkedIn posts, the X threads, the newsletter that went out today. All generated, scored, and published through the pipeline.

LinkedIn — Audit Post

“I opened a new client's Klaviyo account last week and saw the same thing I've seen in the last hundred accounts I've audited...”

Generated by LinkedIn Strategist → Scored 8.2/10 by Chief Strategist → Published via Typefully

Newsletter — The Retention Breakdown

“What I find in the first 72 hours”

5 custom HTML graphics, full flow breakdown, published to 116 subscribers via Klaviyo at 10 AM local time

The Compounding Effect
Preflight checklist — Every correction Raymond makes becomes a permanent rule. The system never makes the same mistake twice.
Gold standards — Approved posts become the voice reference for all future content. Quality compounds.
Voice markers — 9 for LinkedIn, 12 for X. Each platform has its own personality enforced by its own scoring agent.
Getting Started

How to use Claude.
Start here.

Two tools, one mindset shift.

Claude.ai
Start here — Everyone
  • Browser-based, no setup
  • Great for writing, research, analysis
  • Upload files, paste data, ask questions
  • Use Projects for persistent context
Claude Code
Advanced — For building systems
  • Terminal-based, connects to your tools
  • Reads/writes files, runs commands
  • MCP integrations (Klaviyo, Typefully, etc.)
  • Memory across sessions — learns your patterns
The Mindset Shift
Don't
  • "Write me a marketing email"
  • "Give me 10 subject lines"
  • "Summarize this article"
Do
  • "Here's our brand voice. Here's the goal. Here's an example of what good looks like. Write one."
  • "I'll give you feedback. Learn from it."
  • "Build me a system that does this every week."
Your Role

What you can do
starting this week.

Pick one task you do every week. Try doing it with Claude.

Copywriters
  • Show Claude what good looks like — paste gold-standard emails and flows
  • Get first drafts that match voice, then refine
  • Build a reference library Claude can learn from over time
  • Make it client-agnostic — same quality bar, any brand
Implementers
  • Client audit prep — paste Klaviyo data, get a structured analysis
  • Flow architecture — describe the logic, get a build plan
  • Campaign performance reports — upload data, get insights and recommendations
  • Segmentation analysis — find opportunities in client data
Designers
  • Show Claude what good looks like — reference your best email designs
  • Design system documentation — auto-generate from existing templates
  • Asset organization — batch rename, categorize, create spec sheets
  • Build a design reference library that scales across clients
Everyone
  • Meeting prep — "Here's a transcript, give me action items"
  • SOPs — turn your process into a documented playbook
  • Client comms — draft professional emails from rough notes
  • Data synthesis — turn spreadsheets into insights in seconds
  • Download Wispr Flow — voice-to-text is way faster than typing prompts
Roadmap

Where 11 Agency
is going with AI.

The content system is phase one. Here's what's next.

Organic Content Pipelinelive
LinkedIn + X + Newsletter — full generation, scoring, and publishing
Engagement Systemlive
Automated commenting on target accounts + reply management
Social Graphics Pipelinebuilding
HTML/React → PNG rendering for LinkedIn and X visual content
Paid Ads Systemplanned
Meta + Google creative strategy, variant generation, performance analysis
YouTube + Shortsplanned
Long-form and short-form video content system
Client-Facing AI Toolsplanned
Automated audits, reporting dashboards, flow architecture generation
The Moat

The tools are available to everyone. The systems aren't. Anyone can access Claude. The advantage is knowing how to orchestrate it — what to build, how to score quality, how to make it compound. That's what I'm building. Every week the system runs, it gets better. Every correction becomes a permanent rule. That gap widens over time.