E1 · Checklist applied
30 minRun your page through the 16 items and note yes or no on each one. Fix the ones that block.
Done when: the 15 blocking items are marked "yes", with what was fixed noted next to them.
Lesson 8 · 20 min
With AI, publishing a thousand pages costs almost nothing. That's exactly why Google targets scaled content: many pages made to manipulate rankings, "no matter how they are created." This lesson shows the shortcuts going around, why each one is a risk, and the right path: the 16-item checklist and a cadence that fits within human review.
Start the lesson
Goal: recognize scaled content and link schemes, and run every page through the 16-item checklist before publishing.
Tap each block.
The same spam policies have applied to Google's AI answers since 05/08/2026.
Click each one.
Using AI is not prohibited. Google's guidance on AI-generated content accepts AI for researching and structuring original content that a person writes or validates.
The scaled content policy lists: lots of generated pages with no value, scraping results, stitching together content from other pages without adding anything, translating or swapping synonyms to disguise it, and creating several sites to hide the scale.
Google clarified that the spam policies also apply to generative AI answers. And, since 04/13/2026, spam reports can be used for manual action.
"First place in 60 seconds" is usually a keyword with no competition, cherry-picked proof, and an unstable result. Don't promise it, don't buy it.
From inemaSEO, adapted to the 2026 yardstick on items 1 and 7. Run each page through it before publishing.
| # | Check | Blocks? | |---|---|---| | 1 | Direct answer at the start, no "in this article we will…" (clarity for the reader) | yes | | 2 | ≥ 1 verifiable piece of first-hand data, with a link to the evidence | yes | | 3 | No number or quote without a source; nothing made up by the AI | yes | | 4 | Does not cannibalize: the main keyword does not belong to another page of yours | yes | | 5 | Title 50–60 characters with the keyword; description 140–155; only one H1 | yes | | 6 | Internal links: pillar page + ≥ 2 sibling pages + ≥ 2 catalog items; all 200 | yes | | 7 | Valid JSON-LD that matches the visible text; FAQ only with ≥ 3 real questions, without expecting a rich result (ended on 05/07/2026) | yes | | 8 | Real update date | yes | | 9 | Human read: has an opinion and "what went wrong"; sounds like the author | yes | | 10 | Swap test: swapped the keyword and the paragraph still holds? Rewrite it | yes | | 11 | Cadence: ≤ 3 new pages per week in the same cluster | yes | | 12 | Translation only with reciprocal hreflang | yes | | 13 | Author block with a link to the author page (sameAs) | yes | | 14 | Link ladder done: 2–3 older pages pointing to the new one, in the same commit | yes | | 15 | Post-publication: URL 200, title, canonical and JSON-LD present; search engines notified | yes | | 16 | Own images and diagrams when they help (not stock images) | no |
Cluster: <name> Week 1: <page> · <page> (≤ 3) Week 2: <page> · ladder from the previous ones Week 3: <page> · <page> Week 4: review: what got indexed? (course 3) Human review: <who> · <10–20 min per page>
The original checklist asks for an answer in the first 60 words "in the format AI search engines cite" and FAQ markup for the rich result. By the 2026 yardstick: a short answer is clarity, with no source backing it up as an AI factor, and the FAQ rich result stopped appearing on 05/07/2026.
From the method's Plan B: a general drop after publishing a lot is a sign of scaled content. The response is to stop, audit with the checklist, merge or remove the weak pages, and lower the cadence.
Fictional conversation for practice. Pick the answer you would give; the comment appears right after.
Choose your answer. The comment appears right after, and the conversation continues based on your choice.
Use the question page from lesson 4.
Run your page through the 16 items and note yes or no on each one. Fix the ones that block.
Done when: the 15 blocking items are marked "yes", with what was fixed noted next to them.
Build the 4-week plan for a cluster from lesson 3, with at most 3 new pages per week and the name of who reviews.
Done when: the plan has all 4 weeks, none goes over 3 pages in the cluster, and week 4 is for review.
The explanation shows up after your answer.
Question 1 of 4
What does the scaled content policy target?skill
Question 2 of 4
What does "humanizing the text so the detector doesn't catch it" fix?
Question 3 of 4
What is Google's Indexing API actually for?
Question 4 of 4
Your site dropped after you published a lot in a short time. First step?
Choose an option. The explanation appears after your attempt.
Four steps and the exercises for each. Tap an exercise to see its "done when".
entity and name collision
2 exercises
Person and Organization with a fixed @id on every page.
View the lessonterritory and score (lessons 2 and 3)
2 exercises
Map of owner pages and 10 scored terms.
View the lessonpillar page, listing pages, links, news (lessons 4 to 7)
2 exercises
First question page, listing pages with the check, link ladder built.
View the lessonchecklist and cadence
2 exercises
Checklist of 16 applied and 4-week plan.
View the lessonBefore you write in the brand's name, say who it is. The collision test becomes the baseline for course 4.
Requires the entity. The territory says where each page lives; the score says which ones are worth it.
Requires the content plan. These are lessons 4 to 7; the map shows the exercises from lesson 4, and the plan below lists all of them.
This applies to every page from now on: nothing goes live without passing all 16 items.
Who, content plan, pages and guard. Only move up a step when the "done when" of the previous one is true.
After the content comes course 3: measuring what worked, separating your own brand.
Pick a start date and how much time you have per week.