In partnership with
AI Publishing on Amazon KDP: What Actually Works After the Cleanup
The opportunity didn't disappear. It just stopped being easy.
In late 2022 and through most of 2023, a particular kind of excitement swept through online creator communities. The message was simple and seductive: use AI to write books in hours, publish dozens of them on Amazon KDP and collect passive royalties while you sleep. YouTube tutorials promised five-figure months. Reddit threads compared "book factories." The gold rush was loud, fast and crowded.
Amazon noticed.
By mid-2023, the platform had introduced mandatory disclosure requirements for AI-generated content. Shortly after, publishers started reporting dramatic drops in organic visibility for thin, templated books. Reviews flagged content as low quality at higher rates. Some accounts received warnings. Others were shut down entirely.
The people who built their KDP strategy on volume and speed had a rough year. But here is what the noise missed: the underlying opportunity on KDP was never about AI itself. It was always about finding underserved readers in specific niches and giving them something genuinely useful. That opportunity is still there. It just requires a different approach now.
What Amazon Actually Changed and Why It Matters
Amazon's response to the content flood was not a blanket ban. It was a filtering system. The platform began requiring authors to disclose AI-generated content during the publishing process. It also updated its quality guidelines to address books that were clearly templated, repetitive or content-thin.
Algorithmically, books with low read-through rates, high return rates, and poor review signals began losing discoverability. This is how the platform has always worked but the baseline threshold for what counts as "good enough" moved upward. Amazon's business model depends on readers trusting that books on its platform are worth buying. When that trust erodes, so does spending.
The practical effect: books that offered genuine structure, clear value, and a coherent reading experience kept performing. Books that were essentially stitched-together AI output with a cover slapped on them disappeared from search results or never found traction to begin with.
The filter was imperfect, as all algorithmic filters are. But its direction was clear: depth and specificity over volume and speed.
The Strategy That Stopped Working
To understand where the market is now, it helps to name exactly what stopped working.
The model that dominated early 2023 looked like this: pick a broad niche, generate a 10,000-word book using an AI tool, create a basic cover in Canva, publish under a pen name, repeat 50 times. The logic was that some percentage of books would catch traction and the math would work in your favor.
The problem with this model was not just Amazon's response to it. The problem was that it was never really a publishing strategy. It was a lottery ticket dressed up as a business. Readers who bought these books and felt disappointed drove down the metrics that Amazon uses to evaluate quality. The model collapsed under its own weight before Amazon even needed to act.
The people still blaming Amazon's policies for their poor results largely built something that was structurally fragile from the start.
What Niche Selection Actually Looks Like Now
The publishers who are earning steadily in 2024 and into 2025 are not chasing broad categories. They are going two or three levels deep into specific problems that real people need solved.
A book on "productivity" competes with thousands of titles and has no clear buyer identity. A book on managing work schedules as a shift nurse with two kids competes with almost nothing and speaks directly to someone who will feel seen by the title alone.
Specificity is not just a marketing tactic on KDP. It is a research and writing discipline. When you know exactly who you are writing for, the table of contents writes itself. The language becomes more precise. The book becomes more useful. And useful books get better reviews, lower return rates, and stronger algorithmic signals.
Scale Your IRL Campaigns Like Digital Ads
Out Of Home advertising has long been effective but hard to scale—until now. AdQuick makes it simple to plan, deploy, and measure campaigns with the same efficiency and insight you expect from online marketing tools.
Marketers agree: OOH is powerful for brand growth, driving new customers, and reinforcing messaging. AdQuick makes it easy, intuitive, and data-driven—so you can treat real-world campaigns like any other digital channel.
Where AI Tools Still Fit Honestly
The disclosure requirements did not make AI tools irrelevant to KDP publishing. They made the role of those tools more honest.
AI is genuinely useful for outlining, for drafting sections quickly, for generating variation in phrasing and for working through structural problems in a manuscript. Writers who treat these tools as a first draft that still requires real human judgment and editing produce books that are competitive. Writers who treat the output as a finished product produce books that underperform.
The meaningful difference is not which tool you use. It is whether a person with actual knowledge of the subject reviewed, restructured and improved what the tool produced. Readers can feel the difference between a book that was shaped by someone who understands the topic and one that was assembled by someone who simply ran a prompt.
There is also a practical argument here: if you know nothing about the subject you are publishing on, your ability to evaluate and improve the AI output is close to zero. That is where quality breaks down.
The Formats That Are Working Right Now
Practical guides and workbooks: Books that give readers a clear process, with exercises or frameworks they can apply immediately, continue to perform well in non-fiction. They have high perceived value and low return rates.
Niche reference books: Condensed, well-organized information on a specific professional or hobbyist topic. These attract buyers who are actively trying to solve a problem, not browsing for entertainment.
Children's books with clear themes: This category was also flooded, but well-illustrated books with coherent storytelling and age-appropriate language still find audiences. The illustration quality is now the real differentiator.
Series over single titles: Publishers who release a series in a niche rather than isolated titles are building compounding visibility. Each book lifts the others in search and in also-bought recommendations.
The Longer Game
What the cleanup on KDP ultimately revealed is that self-publishing has always rewarded the same things traditional publishing rewards: knowing your reader, delivering on your promise and caring about quality as a function of usefulness.
The tools available to independent publishers today are genuinely powerful. Research that once took weeks can be done in hours. Outlining, drafting, cover concepts, book descriptions, keyword research all of it is faster and more accessible than it was five years ago.
But faster access to tools does not change what the reader ultimately needs from a book. They need to feel that the time and money they spent was worth it. That judgment happens one reader at a time and it shows up in the data Amazon watches closely.
The publishers who understand this are not slowing down. They are building with more intention. That distinction is what separates the people still earning on KDP from the ones who left frustrated and concluded the platform no longer works.
The platform works. The shortcut does not.