Selectable
What happens when a reader asks AI for a book like yours — and you're not there.
“I want a book about a reluctant witch. Modern setting. Emotionally grounded. A little funny, but not cozy.”
A reasonable ask by a reader. After all, we know what we like. Except that in this case, the reader is asking AI and it’s AI selecting the list of books in seconds.
Are you on that list?
No? Yes? You don’t know?
This post covers another aspect of the AI Visibility formula to get your book/newsletter positioned to be on that list — Selectable.
The formula we’re talking about:
AI Visibility = Findable + Selectable + Trustable
In my last post, I covered Findable – using the Google tools to make yourself visible to AI. You can find it here. This week, we’re going to cover what you need to do to make it easy for AI to select your books or newsletters when readers ask for recommendations.
Substackers – I have not forgotten you. I’m writing a special post for you.
Your book makes a promise. Does AI know what it is?
Authors — have you ever thought about the promise you make to your readers? Were you even aware you were making one?
When I was writing The Call to Witchery, I had a reader in my head. Someone curled up somewhere, fully inside the story. I made myself a promise: I would write a book that makes her laugh and catch her breath in the same chapter. That the magic would feel real because the emotions driving it are real. That she would close the last page and carry something with her.
That was my internal promise.
The internal promise you made has to become external.
For The Call to Witchery, that translation looked like this:
The emotional promise: magic driven by grief and memory, humor as relief not parody, a heroine who earns her power slowly and at a real cost.
The external language: contemporary fantasy, reluctant witch, emotionally-driven magic, found family, wit alongside danger.
Same promise. Different words. One lives inside the book. The other lives on the outside — in genre labels, trope descriptions, comp titles — where readers can actually find it.
Genre — the first word of the promise
Genre isn’t a box. It’s a contract.
When I chose contemporary fantasy for The Call to Witchery, I was telling my reader something specific: this magic lives in the modern world, not a distant realm. The stakes are grounded. The heroine has a job, routines, a life she’s built carefully — and something is about to upend all of it.
That’s not just a category. It’s the first word of the promise.
Subgenre tightens it further. Contemporary fantasy covers a lot of ground. Adding “modern witches” narrows it — now the reader knows the flavour of magic, the type of world, the kind of story they’re walking into. It’s not epic fantasy. It’s not grimdark. It’s not portal fantasy, even though there’s a portal in the book.
The genre label is doing work before a single page is turned.
The more specific you are, the closer the match for readers. A vague genre label is a vague promise. And a vague promise attracts the wrong readers, sets the wrong expectations, and ends in reviews that feel like they’re describing a different book entirely.
Tropes — the emotional shorthand
Tropes get a bad reputation. Writers resist naming them, as if claiming one means admitting the book is derivative.
It doesn’t. Tropes are shorthand for emotional experience.
When a reader sees “reluctant heroine,” they know they’re going to watch someone resist a destiny and slowly claim it anyway. That arc — the resistance, the forced awakening, the first real ownership of power — is what they’re coming for. It’s what I’m delivering in The Call to Witchery. Abigail doesn’t want the magic. She’s spent years keeping it buried. The story is what happens when that stops being an option.
“Found family” tells a reader there will be loyalty alongside the danger. People who show up. Bonds that form under pressure and hold.
“Predator-charmer villain” tells them the antagonist is going to be genuinely unsettling — not cartoonishly evil, but seductive about it. Harm Slade is a celebrity philanthropist on the outside. What he wants from Abigail, and how far he’ll go to get it, is something else entirely.
Every trope you claim is a micro-promise. If you name it, you have to earn it on the page.
There’s a principle in storytelling called Chekhov’s Gun — if you put something on the wall in act one, it has to go off by act three. Tropes work the same way. If it doesn’t land, it’s not just a disappointment. It’s a broken contract.
Claim what you actually deliver. Not what sounds good in a marketing pitch.
Comp titles — the reader’s reference points
Comparable titles are the ones most authors either skip or get wrong.
Skip them and you’re asking readers to place you with no reference points. They’re working blind.
Get them wrong and you’re setting expectations your book won’t meet. You attract readers who want something you’re not delivering, and lose the ones who would have loved exactly what you wrote.
The goal isn’t to brag. It’s to triangulate.
For The Call to Witchery, my comp triad is: A Discovery of Witches (Harkness) for the modern witch vibe and the sense that magic has weight and consequence. Uprooted (Novik) for reluctant power and magic that runs on emotion rather than rules. The Very Secret Society of Irregular Witches (Mandanna) for warmth and humor alongside real stakes.
Together those three comps describe the experience: emotionally grounded, witty but not cozy, magic that costs something. A reader who loved all three knows immediately whether The Call to Witchery is for them.
That’s the job. Not “I’m as good as these books.” Not “I want these readers.” But: if those worked for you, this probably will too.
Why AI needs to see the promise
AI can’t read your book.
It hasn’t sat with your characters. It doesn’t know that the humor lands exactly when the tension peaks, or that the magic system runs on grief because that’s what the story demanded. It hasn’t felt any of it.
What it can do is read everything around your book.
Your genre label. Your trope descriptions. Your comp titles. Every public-facing piece of language you’ve put out consistently — or inconsistently — across every platform. Not covered here, but equally important: your blurb, your author bio, and your series name and order.
That’s what AI works with when a reader asks for a recommendation. It’s matching the reader’s request against everything it can see. If what it sees is vague, it guesses. If what it sees is contradictory, it picks something clearer instead. If what it sees is missing, you’re not in the conversation at all.
This is why the internal promise has to become external language. Not for marketing. Not for algorithms. Because a reader is out there right now asking for exactly the book you wrote — and the only way AI connects them to you is if the promise is visible enough to match.
Genre, tropes, comp titles aren’t hoops to jump through. They’re how your book speaks when you’re not in the room.
The craft you already have, facing a different direction
This work feels like marketing. I get it.
Genre labels, comp titles, trope lists — it has the texture of a press kit, not a story.
But the skill it requires is one you already have.
You know how to make a promise inside a story. You know how to set up an expectation and pay it off. You know the difference between a detail that earns its place and one that’s just decorating the page.
Selectable is the same skill, facing outward.
Reply with your genre, your top trope, and one or two comp titles. I’ll tell you whether they’re making the same promise.
Next up: Selectable for Substack — what if you only have a newsletter?