Run It on AI
by StructLabs.io
$9.99
For founders and operators of small companies who already use AI every day. ChatGPT tab open, a few prompts that save time, automations that mostly work. And still the company runs through the same three people.
Run It on AI is a working guide to AI tools, the prompting that gets results, and the automations you can ship on Make.com, n8n, Airtable, and the apps you already use, without the founder in the middle. Seven chapters in three acts: Discipline, Work, Leverage.
Three kinds of work, one decision.
Take any task in your business and ask one question: does the value of doing it reset every time, or does it compound? Resetting work, like status emails, file moves, and the report nobody reads, you automate aggressively. Reinforcing work, like sales calls, knowledge bases, and the client-context document, you augment but never replace. Reframing work is judgment, naming, and direction, and you protect it entirely. This is the book's strongest opinion, and Chapter 1 builds everything else on it.
What the book teaches.
Act 1 — Discipline. How language models actually work under the hood. The fluency trap: the model sounds equally certain when it is right and when it is making things up. The model-vs-harness distinction, where identical model weights produced a 25-point swing on the same benchmark inside a different harness. The two levers you actually control: prompting and context. Then the one discipline that prevents most expensive automations, which is questioning the work before you automate it, and the six consequences that follow. Plus an 8-check readiness review that ends in one of four verdicts: Ready, Almost Ready, Not Ready, or MVP It.
Act 2 — Work. Claude as the harness you sit inside: Chat, Cowork, Code modes, memory, projects, MCP connectors. Then the glue layer that connects the harness to the rest of your operation. Make.com for workflow automation, n8n when you need self-hosting, Airtable for operational data, Stacker and Glide for lightweight frontends. The glue test for when a model belongs inside a scenario at all. And the per-tenant isolation pattern that keeps multi-client workflows from bleeding context into each other.
Act 3 — Leverage. Build your own Skills (small folders Claude loads when the work matches), Plugins (versioned bundles you install in any compatible harness), and Sub-agents (specialists your main session spawns for a defined job). All built on the open agentskills.io standard Anthropic released in December 2025 and donated to the Linux Foundation, so the artifact survives whichever harness you switch to next. Then five turn-by-turn playbooks: contract and proposal review, weekly cross-engagement status synthesis, customer-feedback synthesis, board or client meeting prep, and project after-action. Each one comes with setup-once instructions, the exact prompts in sequence, the token cost per run, and the failure mode to watch for.
What you get.
The full book, 7 chapters in 3 acts, in three formats: PDF, a gated web reader at ebook.runitonai.com, and agent-readable Markdown. An Implementation Block closes each of Chapters 1 through 6; Chapter 7 is the playbooks themselves. Appendix C is a catalog of automation ideas sorted by business area and tagged by which of the three kinds of work each one is. Appendix D is a reader-interview prompt for the impatient. Appendix E is the context-document interview the rest of the book leans on.
Two bonuses, separately.
The companion. A live connection between the AI you already use and the ebook — it applies the ebook's disciplines when you ask AI-automation questions, and because the content comes over the connection, you always have the latest version of every chapter. Connecting takes about two minutes and no terminal: in Claude (the free plan is enough), you paste one web address into Settings, then activate with the license key on your receipt. On ChatGPT, you open the Run It on AI Companion GPT link and give it the same key. Codex and Claude Code connect with one config snippet. Your license key is the only credential the companion asks for — never your email — and the companion server is open source, so you can check that claim yourself.
Appendix D, the reader-interview prompt. Give any AI tool you already use the PDF, the Markdown, or a link to the web reader, paste the prompt, answer a few questions, and walk out with the chapter you should read first and one specific next action you can finish in under thirty minutes.
Where the patterns came from.
Property management. The owner used to triage every maintenance request himself, 30 to 60 minutes of inbox sorting each morning. Intake now routes by property and urgency, AI drafts the work order, and a dispatcher approves in one click. The same small team is running 1000+ homes across five communities with no new operations hires in 18 months.
Commercial diving dispatch. The dispatcher was the bottleneck: every job through her head, on her hours. Intake now collects structured fields up front, and the scheduling layer reconciles divers, vessels, and tide windows automatically, surfacing only the exceptions. 140+ commercial diving jobs a month now move through one operator instead of three.
Education portal. The team carried a recurring tax of status emails. A self-serve portal now answers status questions before they reach a person, backed by a context document the team maintains. 300 students a year move through the workflow without generating recurring inbox work.
Forty-one missing statements. The book is also built on the month Ben broke a Make.com scenario for a client and did not notice for two months. Forty-one statements that should have left the building did not. That failure is where the observability discipline in the book comes from: every workflow in the book ships with a metric, a named owner, and a review cadence.
Written by Benjamin Foong, who runs StructLabs.io and built each of the three systems above. The book is the operations playbook from those engagements, with the client specifics removed.
Who it's for.
Founders and operators of small companies who already use AI every day and want the work to stop being fragile. You need more capacity, but hiring isn't the only lever you want to pull. You are willing to think clearly about process before touching tools. You want the system to survive a week where you aren't paying attention.
What this is NOT.
Not a prompt pack. There are prompts inside, but they are subordinate to the operating decisions. If you want a dump of clever prompts, this isn't it.
Not an ML or data-science book. No fine-tuning, no vector databases, no model engineering. If you are shipping AI as a product feature, this is the wrong book.
Not a passive read. Several chapters will tell you to fix the process before you automate it. That's slower than the other ads promise.
Not a guarantee. A $9.99 book cannot guarantee a financial outcome. Anyone who promises that on a book purchase is lying.
Price and refund.
$9.99 flat. No countdown timer, no fake discount theater. Thirty-day refund: reply to your receipt, no questions asked. v1 updates continue as long as the book can reasonably keep being improved.