The Indie Author’s Workflow for Building Book Cover Drafts in Nano Banana
I have published four novels in five years, all of them self-published, all of them in a niche genre that publishers politely tell me has “too small an audience to justify acquisition.” Cozy fantasy mysteries, if you must know. Older detective characters, small towns with magic, the occasional cursed teapot. I make a modest living from them, mostly because I have learned to do every part of the book production process myself except the editing.
The single part of that process that has cost me the most money over the years has been book covers. Not because I am bad at recognizing a good cover. I am very good at recognizing a good cover. I am bad at making one, and bad at briefing other people on how to make one, and the gap between what I imagine and what I have ever been able to commission has cost me thousands of dollars and at least one book release that underperformed for reasons that were clearly visual.
Why Book Covers Matter More Than Indie Authors Want to Admit
Self-published authors talk about a lot of things — writing routines, marketing funnels, advertising budgets, Amazon ranking strategies — but they tend to talk about covers in slightly defensive tones. There is a strain of indie author culture that insists the cover does not really matter as long as the writing is good, and that obsessing over the cover is a procrastination tactic.
I have come around to thinking the opposite. The cover is not the book, but the cover is the thing that decides whether a stranger ever opens the book. On Amazon, on Kobo, on a Pinterest board, on a BookTok video, on a friend’s recommendation — the cover is doing the entire first impression. A good book with a wrong cover sells maybe ten percent of what the same book with a right cover would sell.
I learned this the hard way with my second novel, which had a cover I had cheerfully commissioned from a designer who specialized in literary fiction. The cover was beautiful. It was also completely wrong for a cozy fantasy mystery. It looked like a Booker Prize shortlist entry, not like a comfort read about a retired witch detective. The book did not sell. When I quietly commissioned a new cover from a designer who actually worked in the genre, and relaunched the same book six months later, sales doubled within a quarter.
The cover was the difference. The writing was identical.
The Old Options for Indie Author Covers
Before AI tools became viable for cover drafting, my options were the same ones every indie author rotates through.
Custom-designed covers from professional cover designers run anywhere from two hundred dollars for a basic ebook cover to two thousand or more for a hardcover-paperback-ebook package from a designer in demand. The good ones are worth every cent. The challenge is finding them, briefing them well, and accepting that the first round may not be the one.
Pre-made covers are the budget option. Cover designers produce stock covers for specific genres and sell them through marketplaces like The Book Designers or GoOnWrite. They cost fifty to two hundred dollars, you swap in your title and author name, and you have a cover. They tend to look fine, but they look fine in a way that means the cover communicates “competent generic indie book” rather than “this specific book is about something interesting.” For a series, premade covers are usually a non-starter because the visual continuity across books has to be designed deliberately.
Doing it yourself with Canva or stock photography is the third option. Some authors are talented enough to make this work. Most are not. The market is full of self-designed covers that quietly mark a book as low-effort to readers who have learned to recognize the visual signals.
Where Nano Banana Changed My Brief-Building Process
The shift that actually mattered for me was using AI image tools not as the final cover, but as a way to figure out what cover I wanted before bringing in a human designer.
The first time I tried Nano Banana for this, I was about a month away from launching my third book, the start of a new series. I had been going back and forth with a designer over email, sending Pinterest boards and trying to describe what I wanted, and we were on round four of drafts that did not quite land. Each round took two weeks. The book launch date was pressing.
I sat down on a Saturday afternoon and started generating my own draft covers in Nano Banana. Not to send to readers — to send to the designer. Same protagonist, same small-town setting, same warm-mysterious mood, in maybe twenty different visual treatments. Soft watercolor over photography. Vintage-inspired illustration. Modern flat illustration. Atmospheric photographic composite. Different color palettes for each. Different focal subjects — the protagonist’s silhouette, the teapot, the village square at twilight, the witch’s cottage in fog.
By the end of that afternoon I had a clear sense of what I actually wanted, in a way that two months of Pinterest boards had not given me. I emailed three of the strongest drafts to my designer with notes about what each one was doing right. The next round of designs came back in days instead of weeks and landed almost immediately.
What an Indie Author Actually Wants From This Workflow
The thing I want to make clear, because I think indie authors who try AI for covers sometimes misunderstand this, is that the goal is not necessarily to skip the human designer. The goal is to have a vastly better brief.
Most cover briefing failures happen because authors do not know what they want until they see what they do not want. Designers spend weeks generating options based on vague descriptions, the author rejects them with vague feedback, and the cycle continues until either everyone gives up or one option happens to land.
Nano Banana shortcuts that cycle by letting the author do their own exploration first. By the time you sit down with a real designer, you can show them six concrete visual directions you have already considered and tell them which one feels right. That makes the designer’s job dramatically easier and the resulting cover dramatically more accurate to your vision.
For some authors, especially those with very small budgets or those releasing many books per year, the Nano Banana cover ends up being the final cover. That is a defensible choice for certain genres and certain markets. For others — and I am one of them — the final cover still goes through a human designer who can do typography, layout, and final composition at a level the AI does not yet match.
Either way, the brief-building stage is where the time and money savings really show up.
Series Consistency Without Locking In Too Early
The other place this workflow has paid off for me is series planning.
A book series needs visual continuity. Reader recognition is doing real work — someone sees the cover of book four on a Goodreads list and instantly knows it belongs to the same series as the book they loved last summer. That continuity requires deliberate design across multiple covers, with consistent typography, palette, and compositional structure.
The problem for indie authors is that you usually do not know what your series will need until you are several books in. The first book in a planned trilogy might end up sharing the world with five spinoff novellas, a prequel, and a holiday short story. Designing the cover for book one without thinking about books five and six is normal, but it locks in visual choices that may not serve the later books well.
Nano Banana lets me sketch out cover concepts for books I have not written yet. Before I commission the final cover for book one of a new series, I generate rough concept covers for the next two or three planned books, just to make sure the visual language I am choosing scales across the series. If the cover concept for book one only works because book one happens to be set in winter, and book two is set in summer, I want to know that now and not after I have already paid for the book one cover.
That kind of forward planning used to require either a very expensive design consultation or a leap of faith. Now it requires a couple of hours in Nano Banana and some thoughtful generation.
What Nano Banana Cannot Do for a Book Cover
I want to be precise about the limits.
Typography is still the largest gap. AI tools have gotten better at rendering text, but the actual decisions about font choice, size, kerning, hierarchy, and how the title interacts with the artwork are still better handled in a real design tool by someone who knows what they are doing. Most great book covers succeed or fail on typography more than on artwork, and the typography decisions are not something I would trust an image generator to make.
Trim-aware composition is another. Print books have specific dimensions, bleeds, and spine considerations. A cover that looks beautiful as a square ebook thumbnail may not translate to a paperback with a spine in the middle. Real designers think about this constantly. Image generators do not.
And there is the originality question, which I want to address directly. A book cover is a piece of marketing, but it is also part of the book’s identity. There is something meaningful about a cover that a real human designed specifically for your book, and that meaning is part of what readers respond to, even if they cannot consciously articulate it. For books I care deeply about, I want a human cover. For drafts, sketches, and exploratory work, Nano Banana is excellent.
A Realistic Workflow for an Indie Author
The workflow I have settled into, after about a year of using this approach, looks roughly like this. For a new book project, I spend an afternoon generating sixty to a hundred draft covers in Nano Banana across a wide range of visual directions. I narrow that pool to the five or six strongest, write up notes about why each one works, and bring those to my cover designer as the brief. The designer produces final designs based on those drafts, and we iterate on the typography, layout, and final polish together.
For minor projects — boxed sets, promotional graphics, marketing imagery, social media art — I often use Nano Banana images directly, with light cleanup in Photoshop and typography added in Affinity Publisher.
For audiobook covers, which are square and have different rules than print covers, I use the same brief-building workflow.
For backlist refreshes, where I am updating an old cover to match my current series style, the workflow is mostly AI-generated with light designer touch-up on the final.
The result is that I produce more books per year than I used to, the covers are stronger than they used to be, and the cost per book has dropped enough that I have actually been able to put money into other parts of the publishing process — better editing, better proofreading, better advertising — that used to get squeezed by the cover budget.
For an indie author trying to make a real living from this work, that shift is not subtle. It is the difference between a sustainable career and a hobby.