Featured long read
The Enabler's Playbook
Agencies build things that break when they leave. Strategy shops sell decks that never get built. Here is why we insist on teaching your team to build it themselves.
Resources
The thinking behind our systems.
Frameworks, field notes, and fundamentals — practical, considered, and written in plain language.
Browse by lane
Choose a lane to filter the list.
- Long read
Case study: Building Contractor Co-Pilot
A seven-month build of an AI operating system for construction — the architecture, the line we drew between AI and human judgment, and the patterns we now bring to every Ochre engagement.
- Long read
An AI system, walked through like a building
Most AI projects fail at the foundation, not the frame. A walk through an AI system as if you were touring a build site — foundation to finishes — so the whole thing finally clicks.
- Long read
The 2026 AI Stack we actually use
Models, tooling, and workflow — what we adopt, what we skip, and how we judge the next shiny thing. No affiliate links.
- Fundamental
AI vocabulary without the vendor speak
Context windows, tokens, RAG, agents — plain definitions tied to what you feel on the ground, not the conference keynote.
- Fundamental
What an LLM actually is (and isn't)
An LLM predicts the next word. Why that's both more and less than people think — and what the mechanics mean for what you can actually trust it to do.
- Fundamental
What AI is for, and what it isn't
The line between draft, retrieve, classify, summarize — and decide. Where you draw it is the thing that determines whether AI helps or hurts your operation.
- Fundamental
How AI fails (and why it does not fail like a person)
Hallucinations, brittleness, confident-wrong outputs. Why 'wrong' from a model looks nothing like 'wrong' from an employee — and why your team has to learn to spot it differently.
- Fundamental
Prompts, context, tools — the three levers
Different controls, different costs, different failure modes. How to know which one you're actually reaching for, and why owners conflate them.
- Fundamental
What "context" really costs you
Why the bottleneck is rarely the model — and how to think about memory, retrieval, and governance before you spend another dollar.
- Field note
Field notes from real engagements
Short, dated observations from inside the work — what broke, what held, what surprised us. Coming as we run more engagements.
Coming soon - Organization & systems
Organization context before models
Why "better prompts" fail when roles, rules, and handoffs are fuzzy — and how we think about fixing that before touching a model.
- Organization & systems
What a business actually is
Units of value, exchanges, repeatable patterns. Business stripped to first principles — the if-you-can't-draw-it-on-a-napkin version.
- Organization & systems
Roles, not people
A role outlives a person. Why owners who can't separate the role from the person can't install systems on top — and the cost of leaving that work undone.
- Organization & systems
Source of truth (and what it costs to lack one)
Every fact in your business needs one place it lives. The price you pay when prices live in three spreadsheets, customer history lives in someone's inbox, and labor rates live in someone's head.
- Organization & systems
Workflows: how value moves through a business
Inputs, outputs, handoffs, queues. How to look at your business as a flow rather than a chart — and why the flow view is the one that surfaces what's actually broken.
- Organization & systems
Deciding vs executing
Two activities owners mix up the most. Why conflating them is the single biggest thing that breaks delegation, hiring, and AI adoption.
Ready to see if this fits how you actually work?
Tell us what you are trying to figure out in one or two sentences. We reply personally — no sequences, no one else to route you through.