The question underneath all the others
People arrive here having already decided AI matters. What they cannot find is a first move they would be willing to defend.
These are the opening words of four enquiries, as written:
- “Do I even start to implement AI in our business?”
- “I need AI. I can see how fast things are changing, but I don't know where to begin.”
- “We have things in our business that I'm sure can run better with AI, but I just don't know how to implement them.”
- “I want to implement AI, but I want to do it properly, protect our client data and make sure everything is compliant and safe.”
Not one of them names a technology. None asks what AI is, or whether it works. The gap is not comprehension. It is the distance between believing it applies and knowing what to do on Monday. Every page in this library exists to close some part of that distance, and none of them will explain what AI is.
Start here, depending on what you already know
Pick the description that fits. Reading in order is not necessary and mostly wastes your time.
You know your processes could be better
You have been working on better processes for years. That part is not new, and this library will not try to sell it to you. What is missing is how AI applies to a specific one. How to start takes the process first and the tool second, which is the opposite of how it is usually sold.
You have built software in the business before
Your instinct for what a build costs and how long it takes was formed before the speed of development changed, so it is calibrated to the wrong decade. The gap worth reading is what happens between something that works and something other people depend on: pilot to production.
You have done the reading and know it can all connect
You are not asking whether. You are asking what the architecture looks like and where it breaks. Start with AI on your own data, then visibility across a business.
The words a vendor will use at you
You will be quoted agent, RAG, orchestration, fine-tuning and copilot, sometimes by people using them loosely and sometimes by people using them to mean whatever the product already does.
The glossary is the most useful page here and the least interesting to write. It separates the terms that mean something, the terms that mean two different things depending on who is talking, and the terms that are mostly marketing. Two of them get their own pages because the distinction changes what you buy: what an agent actually is, and how it differs from automation.
What usually goes wrong, before what could go right
Most people arriving here have watched an AI project go badly, or know somebody who did. That makes confident success stories a reason for suspicion, and it makes the failure pages the ones worth reading first.
- Why AI projects fail: five patterns we have watched play out, written as patterns rather than as statistics, because we have not measured our own and will not borrow somebody else's.
- AI governance at your size: the decisions to take before you build, and what has to be written down. Much smaller than the word suggests, and much cheaper before than after.
- Pilot to production: the specific reason a working demonstration tells you less about cost and time than it appears to.
When you want to know where you actually stand
Two of the four enquiries above are the same question: where do I begin. Reading cannot answer that one, because the answer depends on your systems, your data and which process is actually load-bearing for you.
The assessment is the structured version of that conversation. It returns a written report rather than a score, and it is deliberately not a test you can fail. A number out of a hundred would imply a precision the questions cannot support.
If you would rather read first, is my business ready covers the same ground without giving us your details.
What this library will not do
Stated so you can rule us out quickly if it is the wrong fit.
- Explain what AI is. Some of the people reading this have been building things with Claude Code themselves. A page that starts from first principles insults them and helps nobody else.
- Tell you it is straightforward. The building is faster than it used to be. Getting it running safely inside a business that already has customers is not, and any page suggesting otherwise is selling something.
- Quote a figure we did not measure. No efficiency multiplier, no hours-saved number, no percentage borrowed from somebody else's report.
- Compare us to a named competitor. A standing editorial rule, not a diplomatic one.