This is the foundation level of the Master AI Marketing Agency Certification, rebuilt as a master class. It is written for the people who make, place and measure marketing in the United States: strategists, copywriters, designers, video and audio producers, media buyers, social and community managers, creator managers, PR and media relations staff, email marketers, analysts, account leads, and the agency principals who will set AI policy. It assumes you know marketing and nothing about artificial intelligence. A learner who finishes can decide, for one real task, whether and how AI may be used, who owns the result, and how they would know it had gone wrong.
Modules one and two take apart what is sold to marketers as "AI". Module one separates rules, machine learning and generative AI, explains how language and image models produce output, opens up the retrieval, tools and agents inside marketing products, and reads the evidence on AI use for what it measured. No adoption rate for agencies or PR firms is stated, because none exists. Module two covers the failures you will actually meet: invented sources, statistics and quotes, the uneven line between tasks AI helps with and tasks it harms, creative that drifts toward sameness, and the automation bias that lets errors through review.
Module three sorts client confidences and audience data before any tool sees them: where AI tools put data, state privacy thresholds and opt-outs for sale and targeted advertising, health and children's data, tracking pixels, and a decision method for putting any data into any system. Module four is the spine of the level: the section 5 deception and unfairness tests, the FTC's test for when an agency is liable for its own participation, agencies as intermediaries in endorsements and as agents under the reviews rule, shared duties for commercial email, and the agency's duty of honesty to its own clients. Module five labels everything else you will meet, because marketing has no licensing board: member codes, voluntary self-regulation, trade association frameworks and platform advertising policies, and what each one can and cannot do.
Modules six and seven build work a reviewer can check. Module six teaches a nine-part prompt, supplies claims and figures from the substantiation file rather than asking the tool to generate them, and treats instruction-like text hidden in pasted pages and emails as a real risk. Module seven finds the express and implied claims in AI-drafted copy, explains why a reasonable basis must exist before the ad runs, drafts only from an approved claims file, and covers claims about AI itself, disclaimers that work, and competitor challenges.
Modules eight to ten cover creators, creative and provenance. Module eight explains endorsements, material connections, disclosures that are hard to miss, the federal rule on fake and incentivised reviews, virtual creators and AI avatars, and how to check whether an order or rule you have read about is still current. Module nine covers what the Copyright Office will register when AI helped make a work, human authorship after Thaler v. Perlmutter, digital replicas as a matter of state law, replica terms in talent contracts and New York's synthetic performer disclosure law. Module ten teaches what Content Credentials, labels, watermarks and detectors can and cannot show: provenance is a claim about origin, never proof of truth.
Modules eleven to fourteen turn good habits into an operation. Module eleven designs staged, recorded campaign workflows in which a wrong output cannot become a published claim. Module twelve sets risk levels for agents, customer-facing chatbots and AI voice, and separates real consent and disclosure duties from ones that do not exist. Module thirteen reads a vendor's security and privacy claims, handles client data from source to deletion, and sets a small-agency security baseline. Module fourteen writes an agency AI use policy, using the voluntary NIST frameworks as a structure and placing management system standards, disclosure frameworks and the EU AI Act where they belong.
Module fifteen is a lab of four exercises on fictional clients with the errors already planted. Module sixteen is the capstone, an Agency AI Implementation Plan for one low-risk first use, with its controls, vendor evidence, records, a baseline you measured, honest metrics, failure indicators and named stop authority. A printable workbook and fifteen templates ship with the level.
Everything here is professional education. It is not legal, privacy, advertising-compliance, intellectual property or cybersecurity advice, and it does not replace counsel, a platform's own policies or the law that applies to a campaign. Every statement of authority says what kind of authority it is, whom it binds, and when it was checked. Completing the level earns an independent educational certificate from AI Coalition Network with a public verification page. It is not a licence, a platform certification, a professional designation or membership of any association, it carries no professional education hours, and it is not a regulatory approval.