Guides / AI fluency
AI fluency
Why does AI give generic answers, and how do I describe what I want?
Generic in, generic out. An AI writes from averages until you give it three things: the exact outcome you want, the facts it should work from, and the voice it should use. Describe those and the output stops sounding like everyone else's.
Generic in, generic out.
Ask an AI to "write a follow-up email" and you get a follow-up email that could belong to any business on earth. That is not the model being weak; it is the model doing exactly what you asked.
You gave it a general request, so it answered from the average of everything it has read. It has no idea what you sell, what you charge, or how you talk, until you tell it.
Describe three things.
- The outcome. What does done look like? "A 3-sentence reply that answers the shipping question and offers the faster option" beats "reply to this." Say the shape, the length, and the goal.
- The facts. Hand over the real numbers, rules, and context: your actual lead time, your real return policy, the specific order. Without them the model guesses, and a confident guess about your business is the whole problem.
- The voice. Show, do not describe. Paste a real email you were proud of and say "match this tone." One example teaches more than ten adjectives.
A brief beats a prompt.
The unhelpful version: "write a review request." The version that gets a usable draft: "Ask this customer for a Google review. They bought a winch kit 2 weeks ago and it shipped on time.
Keep it under 4 sentences, warm but not gushing, no discount offer. Match the tone of this email I sent last month: [paste]." Same tool, completely different output, and the difference is the brief, not the prompt trick.
The facts are the hard part, so write them down once.
Retyping your policies and voice into every chat gets old fast. The durable fix is to keep them as plain files the AI reads before it drafts anything, a business knowledge base. Then describing the outcome and pointing at the files is the whole brief, and the facts come along automatically.
Go deeper: a good brief has three layers, the outcome, the method, and the manner.
Description is one of four skills in the AI Fluency Framework (Dakan, Feller, and Anthropic, CC BY-NC-SA), which we adapt in our guide to AI fluency for owner-operators. The examples here are our own.
Apply it
The facts an AI needs are the same ones we write down first.
Month one turns your rules and voice into files any AI can read. What you get · Book a discovery call