Skip to content
Koit Academy

Digital Marketing

AI for digital marketing, organised by the job it does

Three guides, 626 pages and 340 numbered prompts. None of them is about a model. Each is about a job.

What each of these three books does

Pick by outcome, not by tool

The other half of the library teaches Claude, ChatGPT and Gemini. These three assume you can already get a usable answer out of one of them, and spend their pages on the work itself. All three are model-agnostic, so the question is not which AI you use — it is what you are trying to ship this week.

If the job is publishing

AI for Content Creators & Social Media is a production system, not a tour. Thirty-two chapters take one idea through hooks, short- and long-form scripts, captions per platform, thumbnails, and a repurposing engine that turns one asset into a month of posts. It ends where most books start: analytics, delegation and brand deals. The 100-prompt creator library is Appendix A.

If the job is running a business

AI for Freelancers & Small Business follows the money rather than the software. Thirty-five chapters on qualifying leads, discovery calls, proposals, holding your price, client email, onboarding, invoicing, chasing late payment and writing the SOPs that let you hire. It is the book for people who are the whole company.

If the job is spending a budget

AI for Google Ads is the most specialised book in the catalogue and the only one that assumes real expertise. Thirty-eight chapters on responsive search ads at scale, search-term mining, negative keyword architecture, account audits, and Google Ads Scripts written with AI — plus ten production scripts and a validation step at every point where a model would otherwise be trusted with arithmetic.

What all three share

One working pattern, rebuilt in each book's own vocabulary: write a reusable context file about your business once — the Brand Brief, the Business Brain, the Account Context Pack — then constrain the model hard, then verify what comes back rather than trusting it. Each book also ends with a numbered prompt library, so a prompt that worked in March is still findable in October: 100 for creators, B-001 to B-120 for freelancers, G-001 to G-120 for Google Ads.

All 3 guides

The Digital Marketing shelf

Side by side

Which job each book does

GuidePagesWho it's forWhat you'll be able to do
AI for Content Creators & Social Media 2026217Creators and social teams shipping on a scheduleRun a repeatable week: ideas, scripts, captions, thumbnails, repurposing
AI for Freelancers & Small Business 2026189Freelancers and owner-operators with no back officeWin work and run admin: proposals, pricing, client email, invoicing, SOPs
AI for Google Ads 2026220Agencies, in-house teams and anyone spending real budgetOperate accounts at scale: ad copy, search terms, negatives, audits, scripts

Choosing between them

Five questions worth answering first

Do I need the AI Tools guides before these?
No. Each of these three starts with its own short setup section covering the prompting you need for that job, then gets on with the work. The tool guides go deeper on the software itself — useful later, not a prerequisite.
Which one if I run a service business rather than make content?
AI for Freelancers & Small Business. Its 35 chapters follow the shape of the work rather than the shape of the software: qualifying leads, proposals, pricing conversations, client email, invoicing, chasing late payment, and the SOPs that let you hire. The 120-prompt library is numbered B-001 to B-120 by the same job categories.
Is the Google Ads book for beginners?
No, and it says so. It assumes you already run accounts and know what a search terms report is. It is the most specialised book in the catalogue: 38 chapters, 10 production scripts and a prompt library numbered G-001 to G-120, written for people explaining cost-per-lead changes to clients.
Do these three overlap?
Barely. They share a way of working — build a reusable context file, constrain the model, verify the output — and each rebuilds it in its own terms in its opening part. After that they diverge completely, because writing a short-form script, quoting a project and mining negative keywords have nothing in common.
Will these work if my team uses a different AI model?
Yes. All three are written model-agnostic across Claude, ChatGPT and Gemini, and say which of the three tends to be strongest for a given task rather than assuming one. The prompts are plain text and carry no product-specific syntax.

Further reading

From the blog

Also in the library

Want the tools themselves rather than the jobs they do?

Seven guides on Claude, ChatGPT and Gemini, prompt engineering, image generation and a 200-template prompt library.

Browse AI Tools →See all ebooks