The guide library
Eighteen evergreen guides, written to be read once and used for years. The first five form a path — read them in order. The rest are filed by theme; take what the week demands.
start here read in order
01 What an LLM actually is (and isn't) The single most useful mental model for working with AI. What prediction means, where the knowledge lives, and why fluent never means true. 02 Prompting fundamentals that outlive model releases Context, constraints, and examples — the three levers that worked in 2023 and still work now. Everything else is fashion. 03 How to choose AI tools for a small business A decision process, not a product list: what to pilot, what to ignore, and the three questions that disqualify most vendors quickly. 04 The solo operator's AI stack What one person actually needs to run research, writing, support, and bookkeeping with AI — and where paying more buys nothing. 05 What AI still gets wrong (limits, hallucinations, when not to use it) The honest failure catalog: where models break, why they break there, and the tasks you should still do by hand in 2026. working with ai
06 AI and your customer data: a privacy baseline A working rule set for what you can paste into AI tools, what you never should, and the vendor settings worth checking before trouble finds you. 07 The verification habit: checking AI work before it ships Fluent output is not checked output. Five fast verification moves that catch most AI mistakes before a customer, a client, or a regulator does. 08 Writing with AI without sounding like a robot AI drafts read fluent and generic. A practical workflow for keeping your voice: feed it samples, draft in pieces, and edit for the tells readers notice. 09 AI in customer support without losing the customer Where AI genuinely helps support, where it quietly burns trust, and the escalation rules that keep an automated front door from becoming a wall. running the business
10 Automate the boring parts: a starter playbook How to pick the first office tasks worth automating with AI, write the recipe, keep a human checkpoint, and know when to stop adding machinery. 11 Brief an AI like you would brief a freelancer The fastest way to better AI output is a better briefing: context pack, constraints, examples, and a clear definition of done, just like hiring out work. 12 Meeting notes by AI: trust, but verify AI meeting summaries are genuinely useful and subtly wrong in patterned ways. Consent basics, a two-minute check, and where the action items hide. 13 Spreadsheets and AI: real help, one sharp edge AI is excellent at writing formulas and explaining spreadsheets, and unreliable at doing arithmetic on your data. Where each mode is safe to use. judgment & safety
14 AI vendor lock-in: check the exit before you enter AI tools accumulate your prompts, documents, and workflows fast. The exit-ramp checklist: what to export, what to keep outside, what lock-in costs. 15 Coding agents for non-developers: small wins, safely You can now get working scripts and small tools without being a programmer. What coding agents do well, the guardrails, and when to hire a human. 16 AI images for your business: what is safe to use Where AI-generated images are safe for commercial use, where copyright and disclosure get murky, and the checks before one lands on your homepage. 17 Stay current on AI in 30 minutes a week A weekly routine for keeping up with AI without drowning: what to read, what to ignore, and the one question that filters everything.