Field notes.
From running agent fleets in production. Patterns that keep working, patterns that broke, and the honest edges of the method.
- agent-patternsmethodology
What Is Agentic Marketing?
Agentic marketing deploys autonomous AI agents that plan, execute, and adapt marketing work on a schedule. How it works, what it costs, and when it fits.
Read → - methodologyagent-patterns
What It Means to Own Your AI Agent's Codebase
Owning your AI agent's codebase means you control the prompts, state files, and deployment. Here is what real ownership looks like and how to verify you have it.
Read → - agent-patterns
How Eight AI Agents Coordinate Without a Slack Channel
Armada Works runs eight AI agents with no Slack, no message bus, and no shared database. 175 days of operation proved that git commits and a synthesizer agent keep a founder to one five-minute read a day.
Read → - methodologyagent-patterns
Claude Code Consulting: What Founders Get Wrong
Claude Code consulting is not about hiring a developer who knows the CLI. It is about buying a delivery methodology with operational proof behind it.
Read → - founder-notesagent-patterns
What Claude Code Agents Actually Do All Day
A behind-the-scenes look at the daily output of Armada Works' eight-agent fleet after 170 days of continuous operation. Real schedules, real files, real coordination.
Read → - methodology
Forward-Deployed AI: Why the Best Consultancies Embed
Forward-deployed AI means embedding agents into a client's codebase, not running them remotely. Here is why the best consultancies embed and transfer.
Read → - methodologyagent-patterns
What Is an Agent-First Consultancy?
An agent-first consultancy deploys autonomous AI agents into your codebase instead of billing hours. Learn how it works, who it fits, and how it compares to SaaS tools and agencies.
Read → - agent-patternsmethodology
How the Armada Agent Fleet Actually Coordinates
How eight AI agents share one codebase without collisions. State files, cron schedules, the synthesizer pattern, cross-agent escalation, and what 165 days of operation taught us about coordination at scale.
Read → - methodology
Red Flags When Evaluating an AI Agent Consultancy
Six warning signs that an AI agent consultancy will leave you locked in, overpaying, or holding a system nobody understands. A buyer's checklist from 165 days of fleet operations.
Read → - agent-patterns
Agent Governance After Black Hat 2026
Black Hat 2026 proved that AI agent governance is not optional. Two public incidents and 165 days of fleet operation show why prompt-level rules fail and what technical controls actually work.
Read → - methodology
How an Agent Engagement Actually Works: Week by Week
A walkthrough of what happens during an Armada Works engagement, from the first agent brief to the handoff milestone.
Read → - agent-patterns
What a Production Runbook for an Agent Fleet Looks Like
An AI agent runbook covers monitoring, permissions, rotation, escalation, state-file maintenance, and decommissioning. See a real example from 160+ days of fleet operation.
Read → - agent-patterns
The Security Audit Every AI Agent Fleet Needs
A five-layer security audit for AI agent fleets covering credentials, execution boundaries, inter-agent communication, output isolation, and recovery readiness.
Read → - methodology
Before You Hire an AI Agent Consultancy
A buyer's verification checklist for evaluating AI agent consultancies. What to prepare, what to ask on the first call, and what to verify before signing.
Read → - methodologyagent-patterns
Why AI Roles Fail Without Operational Backing
Most Head of AI hires fail because companies design the role without implementation infrastructure. A pillar guide to AI org-design that actually ships.
Read → - methodology
Is Your Codebase Ready for AI Agents?
A 12-question self-assessment for AI agent readiness. Covers repo structure, CI/CD, secrets management, security governance, monitoring, and team readiness, with operational examples from 140 days of running an eight-agent fleet.
Read → - agent-patternsmethodology
Agentic Marketing for Paid Channels
Where AI agents add real value in paid advertising, where they fail, and how to evaluate ROI claims before signing a contract.
Read → - methodology
AI Agent Strategy for Founders: From Hire to Fleet
Three paths to an AI agent strategy: hire a Head of AI, engage a consultancy, or build in-house. Which fits your team, budget, and timeline.
Read → - agent-patternsmethodology
Claude Agent SDK vs Claude Code: When to Use Each
Compare Claude Agent SDK and Claude Code for client delivery. When each fits, how they differ, and when a consultancy should use both.
Read → - methodologyagent-patterns
Why Fractional AI Implementation Beats a Full-Time Hire
Fractional AI implementation delivers working agents from week one without the org-design failure mode of hiring a Head of AI. Compare cost, timeline, and outcomes.
Read → - methodology
What Your First Week With an AI Agent Fleet Looks Like
A day-by-day walkthrough of what happens when Armada Works deploys agents into your codebase, from first commit to first real output.
Read → - agent-patternsmethodology
Head of AI with No Authority: Why AI Org-Design Fails
Companies hire a Head of AI without giving them budget, team, or operational backing. The role fails because AI strategy cannot work without implementation infrastructure.
Read → - agent-patternsfounder-notes
What Happens When Agents Break
Real incidents from running an eight-agent fleet in production: how failures surface, what the founder actually does, and why the blast radius stays small.
Read → - methodology
How to Evaluate an AI Agent Consultancy
A comprehensive buyer's guide for founders evaluating AI agent consultancies. What to look for, red flags, ROI framework, and a complete evaluation checklist.
Read → - methodology
How to Evaluate AI Agent ROI Before Signing a Contract
A framework for founders evaluating AI agent ROI before committing to a consultancy. What to measure, what to ignore, realistic timelines, and red flags.
Read → - methodology
Why an Agent Fleet Costs More Than Marketing SaaS
DFY AI marketing SaaS platforms charge hundreds per month. An agent fleet consultancy starts at $5,000. Here is what the cost gap actually buys.
Read → - agent-patternsmethodology
Multi-Agent Coordination Without Vendor Lock-in
How to coordinate AI agents across providers without coupling your architecture to any one vendor's session model, memory system, or permission framework.
Read → - agent-patternsmethodology
Agentic Marketing Use Cases
Five agentic marketing use cases that replace manual workflows with autonomous agents. Lead scoring, content, SEO, outbound, and social listening.
Read → - methodology
What to Look for in an AI Agent Implementation Partner
Seven evaluation criteria for founders choosing an AI agent implementation partner, from delivery model and code ownership to security posture and pricing.
Read → - agent-patterns
AI Agent Consultancies for Founders in 2026
A founder-oriented roundup of 10 AI agent consultancies and platforms. What each one does, who they serve, and how they differ.
Read → - agent-patternsmethodology
What a Lead Nurture Agent Actually Does
An AI lead nurture agent automates founder follow-up with scheduled emails, state files, and git commits. When it works, when it doesn't, and what the output looks like.
Read → - agent-patternsmethodology
How to Deploy Claude Code Agents into Your Codebase
A step-by-step guide to deploying Claude Code agents: from prerequisites to fleet architecture. Learn what works, what breaks, and when to bring in help.
Read → - agent-patternsmethodology
Claude Code Scheduled Tasks for Marketing Teams
Learn how to set up Claude Code scheduled tasks to run a marketing agent fleet on your own repo. State files, cron cadences, and practical setup steps.
Read → - founder-notesmethodology
What an Agent Fleet Actually Costs a Pre-Revenue Founder
A founder-voice breakdown of what you spend on an AI agent fleet before you have revenue. Armada Works pricing, infrastructure costs, and the honest math.
Read → - methodologyfounder-notes
Six Questions to Ask Before You Hire an AI Consultancy
The six most common questions founders ask before hiring an AI consultancy, with direct answers on visibility, ownership, handoff, timelines, and how agents differ from ChatGPT.
Read → - agent-patternsmethodology
Why Most AI Agent Projects Stall
AI agent deployments fail for four predictable reasons. Learn the failure modes, what successful teams do differently, and how to tell if your company is ready.
Read → - agent-patternsmethodology
How to Build a Lead Nurture Agent with Claude Code
Step-by-step tutorial for building a lead nurture agent using Claude Code scheduled tasks, state files, and an email API. Practical, buildable, and yours to keep.
Read → - methodology
AI Agent Handoff Checklist for Founders
A step-by-step checklist for taking ownership of your AI agent fleet after a consultancy engagement. Covers repo transfer, scheduled tasks, monitoring, costs, and first-week rhythm.
Read → - methodologyagent-patterns
Agentic Marketing SaaS vs. Consultancy
Agentic marketing SaaS platforms run pre-built agents on vendor infrastructure. Consultancies deploy custom agents into your codebase. Compare both models.
Read → - founder-notes
Three Times an Agent Fleet Is the Wrong Answer
Not every company needs an AI agent fleet. Here are three honest signals that tell us you should not engage Armada Works.
Read → - agent-patterns
How to Audit AI Agent Permissions in 30 Minutes
A step-by-step AI agent permissions audit: review .env.local, scope Supabase and GitHub tokens, configure PreToolUse hooks. Complete it in one sitting.
Read → - agent-patterns
How to Monitor Your AI Agent Fleet in Production
AI agent monitoring is not APM. Learn what to watch for, how daily briefs replace dashboards, and the patterns that catch drift before it compounds.
Read → - methodologyfounder-notes
Agent Fleet vs. Fractional CMO vs. Marketing Agency
How three marketing-resource models compare on cost, speed, ownership, and strategy for a seed-to-Series-A founder making the call.
Read → - agent-patterns
Your AI Coding Agent Can Delete Your Entire Production Stack. Here's How to Stop It.
AI coding agents have full access to your production credentials. A PreToolUse hook blocks unrecoverable commands before they execute.
Read → - methodology
What Happens After the AI Consultancy Leaves
After the AI consultancy leaves, two things always break first. Learn the three daily habits that keep your agent fleet running on its own.
Read → - methodologyagent-patterns
The Operate Tier: What a Managed Agent Fleet Actually Does
A managed AI agent fleet runs daily in your codebase, posting briefs and committing work. Here is what the Operate tier delivers.
Read → - agent-patternsmethodology
Marketing Agents vs. Marketing Automation
Marketing automation handles workflows. Autonomous marketing agents handle decisions. Here is when each approach fits and when you need to upgrade.
Read → - methodologyagent-patterns
Agent Fleet vs. Marketing Agency: What's Different
A side-by-side comparison of autonomous AI agent fleets and traditional marketing agencies, covering cost, speed, ownership, and transparency.
Read → - methodology
AI Agent Consulting for Startups: What Founders Need to Know
AI agent consulting helps startups deploy autonomous agents into their codebase. Learn what it costs, what you own, and when your startup is ready.
Read → - methodology
The Transfer Engagement Model: How Handoffs Work
The transfer engagement model gives founders full ownership of an AI agent fleet in 2 to 4 weeks. Here is how the handoff works at Armada Works.
Read → - agent-patternsmethodology
Build Your Own Marketing Agent Fleet
A marketing agent fleet is a coordinated set of AI agents handling SEO, content, outbound, and sales: running autonomously in your codebase via git.
Read → - methodologyagent-patterns
DFY AI Marketing: What Founders Actually Get
Done-for-you AI marketing means a fleet of autonomous agents running SEO, content, outbound, and sales ops in your codebase: with daily briefs and full ownership.
Read →
