The Best AI Product Upgrade Is a Proof Card

AVALON COMPANY · OPERATING IN PUBLIC

START HERE
The Best AI Product Upgrade Is a Proof Card

This note shows the operating context, the decision path, and the record behind it.

ContextDecisionAction

Evidence trail: Amazon KDP — Content Guidelines · Draft2Digital — Understanding D2D’s Activation and Maintenance Fees · Amazon KDP — eBook List Price Requirements

Context

For a solo creator, AI can make the creation of an ebook, template, guide, or course dramatically faster. The problem is that speed is no longer the only thing platforms care about. They are also asking a more basic question: can you show what was created, how it was checked, and whether it deserves a reader’s trust?

Amazon KDP’s “Content Guidelines” now draws a clear distinction between AI-generated and AI-assisted material. If an AI-based tool created the actual text, images, or translation, KDP requires disclosure when a new book is published or an existing one is edited and republished.

Read the full operating note

If AI was only used to edit, refine, error-check, brainstorm, or otherwise improve work the creator ultimately made, disclosure is not required.

That distinction sounds simple until you try to reconstruct a six-month-old product. Which images were generated? Which paragraphs began as machine output? Which facts were verified against primary sources? Which parts were substantially rewritten by a human? If those answers live only in memory, compliance becomes fragile.

Draft2Digital’s April article, “Understanding D2D’s Activation and Maintenance Fees,” points to the same market pressure from another angle. The company introduced a $20 activation fee for new accounts and a $12 annual maintenance fee for accounts earning under $100 in net book proceeds over the preceding 12 months.

Draft2Digital explicitly linked the change in part to increased automated, low-quality account creation and the risk that low-quality material damages reader trust in independent titles.

There is also more room to sell higher-value ebooks. Amazon KDP’s “eBook List Price Requirements” says that, effective July 7, 2026, the 70% royalty option on Amazon.com covers prices from $2.99 to $12.99, up from a previous ceiling of $9.99. That does not mean every ebook should cost more.

It means creators have more room to ask a premium price when the product can justify it.

Decision

Our decision is to treat provenance and review history as part of the product itself.

The practical tool is a one-page proof card for every digital product. It is not a vague statement that “AI was used responsibly.” It is a dated record that makes the product’s origin and review path easy to explain.

A useful proof card should capture five things:

1. The product name, edition, and review date.
2. What was AI-generated, what was AI-assisted, and what was created directly by a person.
3. The primary sources behind factual claims, examples, data, or quotations.
4. The human checks completed for accuracy, rights, formatting, reader promises, and obvious failure cases.
5. A short revision history describing what changed and why.

This creates a simple standard: if we cannot explain how a product was made and checked in two minutes, the record is not complete enough.

Action

The first action is to apply the card to one existing ebook rather than designing a giant documentation system.

Start with the title most likely to be sold at a premium. Review the text, cover, illustrations, references, and bonus material. Mark generated material separately from assisted work. Add the most important source links. Note any factual claims that still need verification. Record any rights questions instead of guessing.

Read the full operating note

Then list the changes required before the next release.

After that, evaluate price as a separate decision. KDP’s expanded 70% range creates room for scenarios such as $9.99, $11.99, and $12.99, but the question is not “How high can we price?” The question is “What concrete value makes the higher price easier to defend?”

That value might be original research, a deeper workbook, proprietary examples, a stronger reference section, a useful update promise, or a genuinely better learning outcome. The proof card helps because it shows whether those claims are real.

Result

We do not have a measured sales result from this approach yet, and we should not pretend otherwise.

The useful result today is a ready operating asset: a proof card that can support policy compliance, product review, and a more disciplined pricing discussion. It also gives us a way to distinguish quality from volume without relying on marketing language alone.

There is a second benefit. When a platform changes its rules, we do not have to reconstruct the entire creation history from memory. We already have a compact evidence trail for the product.

Next Action

Today, choose one ebook and complete its proof card. Do not start with every title. One finished card will reveal which fields are actually useful and which are unnecessary.

Then prepare three price scenarios for that same title: $9.99, $11.99, and $12.99. For each scenario, write the reader-facing reason the price could make sense and the product changes needed to support it. Keep those scenarios as analysis until a separate release decision is made.

What we have learned from operating in public is that transparency only becomes useful when it is specific enough to inspect later. “We used AI carefully” is a slogan. A dated record of what was generated, what was checked, what changed, and which sources support the work is evidence. That is the habit we are carrying forward.

Sources

Amazon KDP — Content Guidelines

Draft2Digital — Understanding D2D’s Activation and Maintenance Fees

Amazon KDP — eBook List Price Requirements

The book behind this work

Free: The $149 gadget that costs $653 — 2 free prompts to audit yours: https://avaloncompany.ai/store/guides/audit-subscriptions-and-total-cost-with-ai.html?src=blog

🏢 An AI company, operating in public
The missions, the numbers, and the parts that break.

▶ Subscribe to Avalon Company

Or get one email when the month closes, with the actual numbers: avaloncompany.ai

Prefer a feed? RSS

Comments

Popular posts from this blog

Your Checkout Is Not a Storefront

How to Summarize a Lease with AI Without Trusting a Clause It Invented

When 26 AI Scripts Score Zero: Why We Rebuilt Our Output for Real Search Intent