The Self Publishing Show SPS-441: AI Can Fix Your Ads James Blatch & Cissy Mecca
Apr 2, 2026
They explore using AI to analyze Facebook ads, sales and reader engagement data in seconds. Practical demos show AI auditing KDP income and BookBub performance to spot best price points and formats. They discuss AI tools for backend admin, reporting and decision-making. The conversation touches on AI translation limits and why detection of AI writing remains unreliable.
AI Snips
Chapters
Transcript
Episode notes
AI Detection Is Unreliable And Unsustainable
- AI detection is unreliable and likely impossible to police at scale for books.
- James argues stylistic signals (like metaphors or punctuation) can't definitively prove AI use and market preference will decide the outcome.
Human Input Determines AI Writing Quality
- AI can mimic human writing but lacks true human complexity, so human input remains the differentiator.
- James suggests the more human effort you put into AI-assisted work, the more likely it will engage readers.
Use AI To Turn Raw Sales Data Into Decisions
- Use AI to analyse advertising, sales, and reader-engagement data instead of manual spreadsheets.
- James fed KDP, Amazon Ads and Facebook data into Claude Cowork and got a consultant-level report in minutes showing revenue, ad spend and read-through issues.