How to Use The Data Letter: AI Operations Builds, Templates, and Case Studies

80+ agent builds, implementation guides, and case studies for senior managers, operators, and technical builders putting AI to work

---

The Data Letter is where I build AI systems in public and hand you the working parts. Agents that run a recurring job for your team. Private AI stacks that run on hardware you own. Compliance and ROI documents that get your AI project approved and paid for. Underneath all of it, the work that keeps your data clean enough to trust what AI does with it.

Every implementation guide comes with what you need to rebuild it. Prompts, code, config, screenshots, and the errors you’ll hit along the way.

---

Pick Your Learning Path

1. Put AI on one recurring task this month

Start here if you want AI to take over a task your team does manually every week.

1. Use Claude Code + Notion to Build an AI Agent That Handles Repetitive Knowledge Work Requests. A written playbook plus an agent that drafts requests on its own.

2. DIY AI Ticket Router Template. Manual triage handed to an agent, with the routing logic included.

3. I Built an AI Agent That Sends Me My Numbers Every Monday Morning. A step-by-step n8n build that reads live metrics and reports on a schedule.

   ---

2. Own your AI stack instead of renting it

Start here if you want AI running without sending company data to an outside vendor.

1. Build AI Tools You Control. How to build AI tools nobody can shut off or take away from you, and what it costs to run them.

2. 3 pieces of free software that install a private AI. Your team, your work, your data, $0.

3. n8n. local. A step-by-step build using n8n, Ollama, and Docker, no API keys required.

4. How To Fine-Tune and Deploy an LLM Without an ML Engineer. Your company’s data stays in-house.

   ---

3. Get the build approved and funded

Start here if the blocker is your CFO, your legal team, or IT.

1. Prove AI ROI to Your CFO in 30 Minutes. A method for showing your CFO what an AI project earns back, before you ask for the budget.

2. Enterprise AI Compliance Workflow for U.S. State AI Regulations in 2026. Intake form, screening guide, vendor checklist, documentation workflow.

3. What New U.S. AI Regulations Mean for Enterprise AI Adoption in 2026. A session on what changed for anyone running AI projects.

4. IT Keeps Blocking Database Access. Here’s the Strategy That Gets It Approved. How to stop getting locked out of your own data.

   ---

4. Keep the systems underneath from breaking

Start here if you’re the one responding to incidents.

1. Drivers of Inevitable Evolution. How a temporary script becomes business-critical, and how to manage that.

2. When Your Tests Pass But Your Data Fails. Testing code isn’t enough. Test your data too.

3. Data Contracts in 5 Minutes. Stop upstream changes from breaking your models.

4. A Proactive Framework for Reliable Data. A Data Quality Index tells you when you’re on fire. Here’s how to fireproof the building.

Full Library by Topic

AI Agents and Workflow Builds

Systems you can copy this week.

• Use Claude Code + Notion to Build an AI Agent That Handles Repetitive Knowledge Work Requests

• DIY AI Ticket Router Template

• I Built an AI Lead Scoring System That Reads My Inbox and Writes to HubSpot. Here’s How.

• I Built an AI Agent That Sends Me My Numbers Every Monday Morning.

• I Just Presented to 20,000+ People. Here’s the Claude System I Showed Them.

• AI Agent Starter Kit

• Build Your Own AI Agent Before Google Ships You Theirs

• Production Hell of AI Agents

• Platform Wars Come for AI Agents

• Build the AI Scoring System Samsung’s 260,000 Employees Aren’t Getting

  ---

Private and Local AI

Models on hardware you own, with your data staying put.

• 3 pieces of free software that install a private AI

• n8n. local.

• Build Your Own Local AI Stack: A Session on Models, Hardware, and Quantization

• How To Fine-Tune and Deploy an LLM Without an ML Engineer

• Build AI Tools You Control

• LLM Fine-Tuning on a Budget

  ---

AI Strategy, ROI, and Regulation

What leadership asks for before they sign off.

• Prove AI ROI to Your CFO in 30 Minutes

• Enterprise AI Compliance Workflow for U.S. State AI Regulations in 2026

• What New U.S. AI Regulations Mean for Enterprise AI Adoption in 2026

• Samsung Gave 260,000 Employees AI. Few Will Use It Well.

• Pinterest Is Turning Taste Data Into an AI Powered Ad and Performance Engine

• Microsoft, AWS, the NFL, and Embedded AI

• NVIDIA and AI Inference Economics in 2026

• LLM Inference Costs

• AI Infrastructure Engineering: Driving JPMorgan’s $2 Billion Return (Part 2)

• Operating Leaders: How JP Morgan Chase Showed the Rest of Us What AI Transformations Can Achieve (Part One)

• Your Company Doesn’t Do Machine Learning Operations. It Does Theater.

• Machine Learning Reality Gap: A Practical Maturity Assessment

Model Choice, Prompting, RAG, and Fine-Tuning

How to pick what you build on, and how to keep it working.

• Prompt Engineering.

• Build a RAG System with NotebookLM in Under an Hour

• RAG in Enterprise AI: Why Companies Get It Wrong and How to Build It Right

• Choosing Between Fine-Tuning, RAG, and Prompt Engineering: A $10K Decision Guide

• Pick an LLM You Won’t Regret. How to Read AI Model Pricing

• Your AI Workflow Will Break on the Next Model Update

• Vector Database Guide

• Implementation Blueprint for Vector Database: Pinecone vs Weaviate vs Chroma

• DeepSeek for Data Analysis

• Code w/ Claude

  ---

Pipelines, Platforms, and Engineering

Plumbing that decides whether the AI layer survives contact with your team.

• Build Your First AI Data Pipeline in Python: From Raw CSV to Predictions

• From Manual Export to Automated Pipeline: A Practical Playbook

• Trapped in the Export Loop

• 5 BigQuery Features That Changed How I Write SQL

• Databricks vs Snowflake vs Fabric

• dbt vs. Dataform: Which Should You Choose in 2026?

• Building a Resilient Data Factory: Data Pipeline Design Patterns That Scale

• Your First Breeze: A Beginner’s Guide to Airflow DAGs

• Drivers of Inevitable Evolution

• When Your Tests Pass But Your Data Fails

• Data Contracts in 5 Minutes

• How Netflix Does Data Reliability

• MLOps on a $50 Monthly Budget

• A Small Data Manifesto

• ‘It Works on My Laptop’ Is Not a Product Strategy

• IT Keeps Blocking Database Access. Here’s the Strategy That Gets It Approved.

• ‘Stop Learning SQL and Python’

Data Quality, Governance, and Monitoring

Feed an agent bad data and it repeats the mistake faster than your team can.

• Who Owns Data Quality, Anyway?

• Data Quality is a Spectrum, Not a Switch

• A Proactive Framework for Reliable Data

• Bad Data’s Hidden Toll: How to Calculate Your Data Debt

• Finding Your $100,000 Query

• Tool Review: Soda Core vs. Great Expectations

• How to Detect Model Drift When You Can’t Measure Performance

• Advanced Model Drift Detection

• From Data Lineage to Data Observability

• Understanding Data Lineage

• Data Privacy Laws in 5 Minutes

• DIY Data Catalog Template

• Data Catalog Implementation: Why 60,000 Tables Break Your Metadata Strategy

• Your Data Catalog is a Ghost Town: Playbook I Used to Fix This for Mars, Inc.

• Single Source of Truth is a Myth

• Night of the Living Dead DAGs

• Your Dashboard Looks Great and Changes Nothing

AI Failure Case Studies

What broke, what it cost, and the checks that catch it next time.

• My AI Gave Me Fake Data: Here’s How to Catch It If It Happens to You

• AI Was Tasked with Growth. It Optimized Away Best Customers

• Our AI Convinced Us We Had a Million New Users (We Didn’t)

• $12M Snowblower Crisis: A Case Study in Why AI Rigor Outlasts AI Hype

• What Happened When a SaaS Company Discovered $210K in Wasted Ad Spend

• How Reinforcement Learning Fixed a $4.8M Promotion Problem for a Global CPG Giant

• Year in Review: Top 3 Data Failures of 2025

---

What Gets Published

Free articles cover what’s changing in AI operations and what it means for how your team works.

Paid implementation guides with prompts, code, config files, templates, and the errors you’ll hit.

Live sessions run every week. Some weeks I build in front of you, other weeks I break down what changed in AI and take questions live. Recordings posted to the archive afterward.

---

What Paid Subscribers Get

• AI Ops Roadmap. A 30-day plan that takes one task your team repeats every week and turns it into a workflow they run without you.

• Every implementation guide, in full. Templates, scripts, prompts, and config you can copy into your own stack.

• Full archive access. Every build ever published, unlocked the moment you subscribe.

• Live builds and replays. Mid-week sessions plus the recordings.

$10/month or $100/year.