All topics
AI for operators
AI Agent and Claude Code Workflows for Operators
These articles cover building and running AI agents in a business: design patterns, guardrails, testing, scheduling, costs, and build logs of automations and skills.
Three good places to begin.
-
AI Agent Workflows: How to Chain Agents Into Multi-Step Business Automations
Learn how to build AI agent workflows that chain multiple agents into multi-step business automations. Patterns, real examples, and error handling for operators.
Read more -
Claude Code Automation: How I Run Four Businesses From a Terminal
Claude Code automation lets operators build skills, connect tools, and schedule agents that run without supervision. Here's how I built mine.
Read more -
MCP Servers: How to Connect AI Agents to Every Tool You Already Run
MCP servers connect your AI agents to the business tools you already use — CRM, calendar, project management, notes. Here's how I set mine up across four ventures.
Read more
All 57 articles, newest first
- AI Agent Documentation: How to Write Runbooks That Keep Your Automations Running Without You AI agent documentation is what separates a real system from tribal knowledge. Here's the runbook framework I use across 30+ agents and four businesses.
- Claude Code Automation: How I Run Four Businesses From a Terminal Claude Code automation lets operators build skills, connect tools, and schedule agents that run without supervision. Here's how I built mine.
- The Skill Was Named After a Model It Didn't Use: An Amazon SBV Video Build Log Build log for an Amazon Sponsored Brands Video skill: 7 AI clips, a 10-second fast-cut, under $5 in generation, and a QA pass that missed garbled label text.
- CLAUDE.md Guide: How to Configure Claude Code to Actually Understand Your Business A CLAUDE.md guide for business operators. Learn how to write the configuration file that gives Claude Code persistent context about your business, workflows, and standards.
- AI Automation Maintenance: How to Keep Your Business Agents Running Without Babysitting Them AI automation maintenance is the routine that keeps 30 agents healthy. The weekly and monthly process I use to catch decay before it costs me money.
- AI Agent Orchestration: How I Coordinate 30 Agents Without a Platform Learn practical AI agent orchestration patterns for operators — supervisor, pipeline, and shared-state coordination using flat files, not enterprise platforms.
- The Skill That Doesn't Do Anything: An Amazon Listing Pipeline Build Log Build log for an orchestrator skill that chains eleven Amazon listing artifacts and four other skills. What broke, what the fidelity gate caught, and what to copy.
- Claude Code Hooks: How to Build Automated Triggers Into Every AI Agent Run Learn how to use Claude Code hooks to build automated triggers, guardrails, and logging into every agent action. Real configs from 30+ production agents.
- AI Meeting Automation: How to Build a System That Handles Prep, Notes, and Follow-Up Without You Build an AI meeting automation system that handles pre-meeting research, captures notes, extracts action items, and sends follow-ups — so meetings stop leaking time and money.
- How to Reduce AI Agent Costs Without Losing Output Quality Practical ways to reduce AI agent costs — model routing, prompt caching, context trimming, and batching — without losing the output quality your business depends on.
- How to Write a CLAUDE.md File That Actually Makes Your AI Agent Useful A practitioner's guide to writing CLAUDE.md files that give Claude Code the context it needs to ship real work — with structure, examples, and common mistakes.
- The Rollout Skill: Cloning a Winning Image Stack Across Four SKUs Without Cloning the Screenshot Border Build log for the Codex skill that replicates an approved Amazon image stack across sibling SKUs. What it extracts, what it refused to claim, and the brand element that turned out to be a screenshot.
- How to Schedule AI Agents: Recurring Automations That Run Your Business on Autopilot Learn how to schedule AI agents for recurring business automation — daily briefings, weekly reviews, and monitoring that runs without you touching a keyboard.
- Claude Code Skills for Business: How to Build Slash Commands That Run Your Operations Claude Code skills turn your best prompts into reusable slash commands anyone can run. Here's how I build, organize, and run skills across four ventures.
- The Infographic Skill That Has to Please Two Readers: A Build Log Build log for a Codex skill that briefs Amazon secondary images for shoppers and OCR at once. The prompt, the constraint that matters, what I'd change, and what it costs.
- How to Build an AI Research Agent for Your Business That Replaces 10 Hours of Weekly Googling Build an AI research agent that does deep business research on demand — competitors, suppliers, markets, opportunities. Real configs and workflows for operators.
- AI Agent Design Patterns: 5 Blueprints for Business Automations That Actually Work Five AI agent design patterns every operator needs. Real blueprints for building business automations that run reliably — from simple transformers to compound accumulators.
- AI Email Automation: How I Built Three Agents That Handle 80% of My Business Email AI email automation cut my inbox time from 90 minutes to 15 minutes a day. Here are the three agents, the prompts behind them, and the mistakes to avoid.
- Consistent AI Output: How to Build Business Agents That Deliver Reliable Results Every Run Get consistent AI output from your business automations. The operator's system for temperature, constraints, validation, and testing that makes agents reliable.
- How to Automate Data Entry With AI Agents Automate data entry with AI agents that handle messy inputs, make judgment calls, and move data between systems. The operator's step-by-step playbook.
- The Skill That Reads a Surface With No Report Behind It A build log: mining Amazon autocomplete in the Codex browser. Why the read-only rule is methodological, not just safe — and the guardrail class none of my other skills needed.
- Automate Business Reports With AI: How I Build Reports That Write Themselves Learn how to automate business reports with AI agents. The exact system I use for weekly revenue, client status, and operations reports across four ventures.
- How to Build AI Agents Faster: The Iteration Playbook That Turns One Good Agent Into Twenty Learn how to build AI agents faster with a structured iteration cycle. Real prompts, testing patterns, and deployment workflows that compress agent development from days to hours.
- AI Daily Workflow: How I Run Four Ventures on 30 Agents and Three Hours of Focused Time Most operators use AI like a search engine. Here's the AI daily workflow I run across four ventures and 30 agents, and what each block of the day actually looks like.
- The Skill That Goes and Gets the Evidence: Building a Read-Only Amazon Review Audit in the Browser A build log for the Amazon review audit skill — stratified star sampling, a read-only boundary on a logged-in session, and why an empty selector is not an empty listing.
- How to Delegate to AI Agents: The Operator's Playbook for Getting Real Work Off Your Plate Learn how to delegate to AI agents with a proven framework covering delegation readiness, 4 autonomy levels, briefing templates, and the trust-building loop.
- AI Agent Testing: How to QA Your Business Automations Before They Cost You a Client A practical AI agent testing framework for operators. Build test cases, catch regressions, and validate your automations before they touch real data or real clients.
- AI Automation Examples: 7 Agents Running My Business Right Now Real AI automation examples from production. Seven agents with actual configs, costs, and time savings running ecommerce and advisory businesses every day.
- I Stopped Asking AI for Hero Image Ideas and Gave It 52 to Choose From A build log: how I replaced generated hero-image concepts with a 52-tactic retrieval library, a selection rule set, and a validator that actually runs.
- How to Future-Proof AI Agents: Build Automations That Survive Every Model Update Most AI agents break every time a model updates. Here's how I build future-proof AI agents across 30+ automations that survive model swaps without rewrites.
- How to Schedule AI Agents: The Operator's Guide to Automations That Run Without You Learn how to schedule AI agents to run on autopilot — cron jobs, event triggers, and the patterns that turn manual automations into overnight operators.
- AI Agent Guardrails: How to Build the Constraint Layer That Makes Your Automations Production-Ready AI agent guardrails are the constraints that keep your business automations reliable. Here's how to write, test, and maintain the rules your agents actually follow.
- Human in the Loop AI: How to Build Agent Automations That Know When to Stop and Ask Learn how to design human in the loop AI agent automations with escalation triggers, autonomy levels, and handoff patterns that protect your business without killing speed.
- I Audited My Own Amazon Creative Skill Library: 11 Automations, 831 Words, and the Word 'Only' in 9 of Them A build log on the Amazon creative automations I actually run. The workflows took an evening. The constraint sentences took longer — and they are the product.
- AI Agent Feedback Loops: How to Build Automations That Actually Improve Over Time Most AI agents never get better. Here's how to build AI agent feedback loops that compound — with the scoring system, correction workflow, and weekly cycle I use across 30+ production automations.
- AI Agent Security: How to Lock Down Your Business Automations Without Slowing Them Down AI agent security matters when your automations touch real business systems. Here's the operator's playbook for permissions, credentials, and audit trails.
- AI Agent Workflows: How to Chain Agents Into Multi-Step Business Automations Learn how to build AI agent workflows that chain multiple agents into multi-step business automations. Patterns, real examples, and error handling for operators.
- How to Monitor AI Agents in Production So Nothing Fails Silently Your AI agents look fine until they don't. Here's how to monitor AI agents in production with alerting and quality checks that catch silent failures early.
- AI Agent Not Working? How to Fix Broken Business Automations Your AI agent stopped working overnight. Here's the operator's playbook for diagnosing and fixing broken AI automations before they cost you real money.
- How to Build an AI Agent for Your Business: The Step-by-Step Operator's Guide Build an AI agent for your business in seven steps — from scoping the job to running it on autopilot. A practitioner's guide with real configs and scoring.
- How to Test AI Agents Before They Touch Your Business A practical playbook for testing AI agents before deploying them to production. Five testing layers, real examples, and the regression loop that makes automations reliable.
- AI Agent Management: How to Run, Review, and Improve Your Business Automations Every Week AI agent management is the discipline of keeping your automations performing. Here's the weekly system I use across 30+ agents and four businesses.
- AI Agent ROI: How to Measure Whether Your Business Automations Are Actually Worth It A practical framework for measuring AI agent ROI across your business automations. Track cost per task, hours reclaimed, and know which agents to keep, improve, or kill.
- The Attribute Fill-Rate Auditor: A Claude Code Build Log I built a Claude Code skill that audits Amazon attribute completeness against the category taxonomy and drafts source-backed fixes. Here's what broke.
- How to Build an AI Chief of Staff That Actually Runs Your Business Operations Step-by-step guide to build an AI chief of staff using Claude Code, MCP servers, and compound memory. Not another chatbot — a real operating layer.
- How to Build an AI Prompt Library for Business That Compounds Over Time Most AI prompt libraries rot in a Notion graveyard. Here's the system I use across 4 ventures and 30+ automations to build a prompt library that gets better every week.
- AI Automation Mistakes: 7 Lessons From Running 30 Agents in Production Most AI automation mistakes don't show up until your agents are running in production. Here are the 7 failures I hit across 30 automations — and the exact fixes.
- AI Agent Costs: What It Actually Costs to Run Your Own Business Automations Most AI agent cost guides quote $50K+ dev fees. Here's what it actually costs when you build automations yourself — real numbers from 30+ agents across 4 ventures.
- MCP Servers: How to Connect AI Agents to Every Tool You Already Run MCP servers connect your AI agents to the business tools you already use — CRM, calendar, project management, notes. Here's how I set mine up across four ventures.
- Vibe Coding for Business: How I Build Custom Operator Tools in Hours, Not Months Vibe coding for business lets operators build custom tools by describing what they want to an AI agent. Here's the workflow I use and what I've shipped.
- How to Build an AI Agent Stack That Actually Runs Your Business An AI agent stack connects your automations, memory, and context into one system that compounds. Here's the exact architecture I run across 4 ventures and 30+ automations.
- I Built a Client's Inventory Brain in a Weekend With Claude Code: A Build Log A build log of a multi-source Amazon inventory and demand portal — SP-API, AWD, 3PL, and Shopify unified in Postgres with Claude Code. The stack, the crosswalk that broke it, and what an operator can copy.
- I Built a Claude Agent That Files My Email Into Structured Notes Every 10 Minutes A build log for an unattended Claude Code + Gmail agent that turns labeled email into structured, cross-linked vault notes. The hard part wasn't writing notes — it was never losing an email.
- The Automation That Watches My Other Automations: A Detect-and-Reverse Build Log I built a Claude Code reply monitor that stops my email sequences from messaging people who already replied. Build log, the prompt, what broke, and the operator pattern.
- How I Automated Our Agency's Daily Intelligence Briefing: 40 Minutes Saved Every Morning, $5/Month in API Costs Build log: using Claude Code CLI, web search, and launchd to auto-generate and email a 10-story daily briefing to the Velocity Sellers team — what broke, what fixed it.
- I Built a Seamless Premium A+ Generator With Claude Code and Gemini — Here's the Build Log A real build log: how I generate five-module seamless Premium A+ pages with Claude Code, Gemini 2.5 Flash Image, and a Pillow seam-enforcement pass. Cost, prompt, and what broke.
- I Built A Claude Code Automation That Turns Every Client Call Into Todoist Tasks — Here's The Build Log Real agentic workflow: Fathom transcripts → action items per attendee → Todoist tasks, fully automated. The prompt, the failures, the cost, the time saved.