Four episodes that frame AI for the small business owner who's heard the marketing and isn't sure where to start. Not the breathless 'AI changes everything' version — the operator-tier read on which parts of the business are actually candidates for automation now, which need to wait, and what 'profit lever' actually means in this context (it's almost always cost reduction or capacity expansion, not new revenue). Format is conversational; the content gives you a specific read on whether your business should adopt AI tools this quarter or whether the time is better spent elsewhere. Written for the owner who wants the honest answer, not the consultant pitch.

The framework for adopting AI in a small business without lighting the operating budget on fire. Where it actually saves time (customer service triage, content drafting, knowledge management, basic data analysis), where the savings are overstated (anything requiring real reasoning, anything where errors compound), and the implementation pattern that works: pilot one workflow, measure, expand only when the first one is stable. The book's contribution is structural — most AI advice is feature-led ('use this tool!'); this is decision-led, walking through which problems actually deserve an AI-shaped solution and which are better solved by simpler means. For the owner running a 1-15 person business who'd rather adopt AI deliberately than reactively.

A week-by-week plan for taking one workflow from 'we should probably try AI' to 'AI is doing this reliably' inside ninety days. Days 1-14: workflow selection (which one, scoped properly). Days 15-30: prompt engineering and tooling setup (the actual building). Days 31-60: pilot and iteration (with the failure-mode catalog you'll need). Days 61-90: hardening and handoff (so it survives once the founder isn't watching). The structure prevents the most common failure pattern: trying to AI-ify five workflows at once, getting partial wins on all of them, and reverting to manual when the energy runs out at week six. One workflow, deliberately.

The honest version of 'is this AI tool worth $40 a month'. Walks through the actual time-savings calculation (against fully loaded labour cost, not headcount cost), the implementation cost most subscription pricing hides (your time tuning the workflow, the cost of a wrong answer in production), and the cost-of-not-adopting framing for cases where competitors will get there first. Not a sales pitch for AI — about half the calculations in the worked examples come back showing the tool isn't worth it for businesses below a certain volume. Useful precisely because it gives the owner the structured way to say no, not just yes.

Seven applications of AI in a small business that aren't obvious from the marketing material — because the marketing material is mostly about content generation and chatbots. The actual leverage tends to be elsewhere: bid optimisation in paid acquisition, automated quality assurance on outbound, structured customer call summarisation, semantic search across internal docs, predictive inventory stocking on top of historical data. Each lever gets a specific implementation note (which tool, what the rough cost is, what 'wins' looks like). Read time about ten minutes. The framing is what makes it useful, not the discovery — most owners already have the building blocks; this names the patterns to plug them into.

Eight email lessons that take an owner from 'should I be doing this?' through to a working pilot. Each lesson contains one specific exercise applied to your actual business, not a generic framework. By session four you've selected a workflow and a tool; by session six the workflow is in production; sessions seven and eight are about hardening it so it survives without constant attention. Built specifically to defeat the failure mode where small business owners take an AI course, finish it, and never deploy a single thing — every lesson forces a small concrete output.

The working prompts for the core small-business AI use cases — not 'write me a blog post' but the structured prompts that produce useful output reliably: customer call summary in the format your CRM expects, weekly metrics narrative from raw numbers, response template generation that matches your brand voice (extracted from existing emails, not invented). Each prompt comes with the input format it expects and the output format it produces. Designed to be pasted into Claude or ChatGPT; tested across both. The value isn't in any single prompt; it's in having a 'standard library' of prompts that produce consistent output, so AI stops being a guess-and-check exercise.

The current state of the small-business AI tool landscape, organised by what they actually do well rather than how they market themselves. Includes the ones worth paying for (and where the cheap or free version is enough), the ones to avoid (named, with reasons), and the integrations that compound — the small set of tools whose value compounds when used together (e.g., automation platform plus AI agent plus CRM trigger). Updated against current pricing and features, not the version that was true eighteen months ago. Skips the listicle pattern of 'fifty-seven AI tools for X' — there are twelve to fifteen that actually matter for a small business, and the document is honest about that.