The skill and its limits
A longer draft needs a claim-by-claim source trail, not invented material to fill space. Practice on two short, complete fictional notes. The exercise deliberately stays small; it does not promise a finished whitepaper or a production timeline.
Practice goal: Separate the observed count, its limited scope and unanswered questions in an outline and paragraph.
Use only the invented facts below. Do not substitute customer lists, CRM exports, private messages, employee information or confidential product plans. Changing names does not establish permission to share real records.
Your complete fictional brief
- Fictional note A: 10 club volunteers were invited to compare two printed notice layouts. Six replied: 4 preferred layout A and 2 preferred layout B.
- Fictional note B: the other 4 invitees did not reply. No reasons, quotations, task timings or later behavior were collected.
- These were voluntary replies from one invented club, not a representative audience study. No sales, reading-speed or conversion result was measured.
Everything needed for the exercise is here. You can work on paper without an AI account, an upload, a purchase or a connection to a marketing system.
Try it and inspect the flawed sample
Write your own answer first. If you choose an external assistant, use only these invented facts. No real audience or customer data belongs in this practice.
Create a three-heading article outline and one short evidence paragraph from notes A and B. Identify the note supporting each factual statement. Keep nonresponses separate from preferences. Do not invent a quote, study, result, web citation or general audience recommendation.
Complete fictional facts:
- Fictional note A: 10 club volunteers were invited to compare two printed notice layouts. Six replied: 4 preferred layout A and 2 preferred layout B.
- Fictional note B: the other 4 invitees did not reply. No reasons, quotations, task timings or later behavior were collected.
- These were voluntary replies from one invented club, not a representative audience study. No sales, reading-speed or conversion result was measured.
Authored flawed sample, not an actual provider response:
Our research proves layout A doubles reading speed. Eight of ten customers loved it. One said, "The best design ever." Every organization should adopt A immediately, according to the National Layout Study.
Feedback and a supported correction
-
Note A records 4 of 6 replies preferring A and 2 preferring B, not 8 of 10 customers. Four nonresponses are not preferences.
-
Note B contains neither quotations nor timings. The quote and doubled reading-speed claim are invented.
-
One voluntary fictional club pilot does not establish a universal recommendation. The named national study is absent.
One supported correction:
Outline: 1. What the club asked (note A). 2. What the replies and nonresponses show (notes A and B). 3. What remains untested (note B and the stated scope).
Evidence paragraph: Of 10 invited club volunteers, 6 replied; 4 preferred layout A and 2 preferred B (note A). The remaining 4 did not reply (note B). No reasons, quotes, task timings or later behavior were collected (note B). This voluntary fictional pilot does not establish a broader audience preference or a performance gain.
Check your answer:
- Preserves all counts: 10 invited, 6 replies, 4 preferring A, 2 preferring B, 4 nonresponses.
- Attaches statements to supplied notes rather than invented citations or quotes.
- Keeps the scope limited and invents no speed, sales, conversion or universal recommendation.
Different wording can pass. Recheck the facts after every tone or length edit. Passing this exercise establishes neither campaign performance nor legal or platform approval.
Sources and your next step
The FTC advertising guide discusses evidence for claims and misleading omissions. Its data guide addresses necessary information and access. NIST describes confidently false output and privacy risks. These sources do not validate this exercise as a campaign or certify a provider.
- Advertising and Marketing on the Internet: Rules of the Road
- Protecting Personal Information: A Guide for Business
- NIST AI 600-1: Generative Artificial Intelligence Profile
The brief, sample and correction are authored fiction. No provider test, human professional review, measured time saving, click rate, conversion result or earned qualification is claimed. Real marketing needs its own current evidence, applicable requirements and organizational approval. Nothing here sends, uploads or publishes a campaign.
Start free Module 1: an account is required; no card is needed. Or read the marketer learning overview.
Frequently asked questions
- Can an expansion prompt supply missing evidence for a long article?
- No evidence is supplied merely by making text longer. For this exercise, mark a missing fact as unknown or remove the unsupported claim; do not turn an invented example into a reported study.
- Can I replace the invented brief with customer or CRM data?
- No. This practice uses supplied fictional facts only. It does not establish permission to share a customer list, private message, employee record or confidential business information with a provider.
- Do I need a paid AI tool for this exercise?
- No. You can write an answer on paper and compare it with the facts, feedback and checking criteria. No model availability, pricing, response limit or provider result is promised.
- Were the samples tested in real campaigns?
- No. They were authored as teaching examples, not collected from a provider or tested with customers. No speed, accuracy, engagement, conversion or sales result is claimed.
- What does the free lesson include?
- Free Module 1 offers general AI learning with an account and no card required. It does not certify marketing competence or approve any real advertisement, message, data upload or campaign.