Seeing the Story: How Frustration Sparked a Revolution in Revision

Q: Who is Will Frasier?

A: I'm a creative. I was a composer and music producer. I’ve released seven albums of my own music. I have written three novels.
Fun fact: when I was 23 I pitched a screenplay to Bruce Willis. It ended up not getting made, but it was a wild ride.

I live in Seattle with my partner Jessica and my three teenage kids, and her two younger children. Outside of work, I’m into reading, art, and human pre-history. My favorite books are The Goldfinch (fiction) and The Dawn of Everything.

Q: What was the moment that sparked the idea for Story Stream?

A: Honestly, it came out of frustration. I was deep into a novel I’d been working on—Darkness in the Desert—and I knew something about it wasn’t working. The story just wasn’t landing the way I wanted it to. So I did what any AI-curious writer might do: I fed it into ChatGPT and asked for help.

The feedback I got was... fine. Polite. Surface-level. Stuff like "show don't tell" or "add back stories for minor characters." Nothing wrong, but nothing useful either. It didn’t see the story the way I saw it. It couldn’t trace the internal logic, the emotional scaffolding—the stuff that actually gives a story weight.

That was the moment I realized if I wanted a tool that understood stories from the inside out—like a writer does—I’d have to build it.

That was the seed for Story Stream. What started as a way to fix my own novel turned into a crash course in editing theory, narrative structure, and the weird art-science of what makes stories work. And in building this thing, I’ve become a much better writer. I can actually see what’s broken now—and more importantly, how to fix it.

Q: As a writer, what frustrations did you face with traditional editing tools?

A: I wanted answers to human questions. Was my villain compelling or just a plot device? Did the pacing drag in the middle? Was I accidentally dumping exposition instead of weaving it in? Did my protagonist’s struggle actually land emotionally?

But most of the tools out there gave me surface-level feedback—stuff like how many adverbs I used, how often I repeated certain words, my dialogue-to-prose ratio. I even saw reports on passive voice frequency and sentence length variance.

It felt like running spellcheck on a symphony. Technically accurate, but completely missing the music.

That matters because writing isn’t about stats—it’s about connection. At the heart of any story is an idea we’re trying to share. And the only way to land that idea is through the craft: character, pacing, structure, emotional rhythm. If those aren’t working, the deeper meaning never makes it through. I didn’t need a grammar coach—I needed something that could see the story under the sentences.

Q: What does it mean for a tool to respect a writer’s voice?

A: Respecting a writer’s voice means recognizing that it’s not just style—for me, it’s the fingerprint of the writer’s perspective, values, and intent. All writers draw from the same toolbox: plot structures, tropes, even turns of phrase. But what makes a piece of writing distinct is the way those tools are used. The author's unique take on the world. What they choose to highlight. What they choose to omit. That’s voice. And it’s not a subtle difference—it’s foundational.

That’s why one of the first things I built into Story Stream was the Author Intentions. It’s just four questions, but after I implemented it, it changed everything about the way the AI was working. It gave it much more nuance and context.

The way it works: writers tell us what they’re trying to do with their story. Are they parodying a genre? Trying on a new style? Subverting expectations on purpose? Writing sci-fi that’s really about the present day? That context is threaded through the entire analysis process. It lets the system anchor its feedback in what the author is actually trying to achieve.

Without that grounding, the tool would end up evaluating every book against the same generic scale. Which in my humble opinion, is a fast track to mediocrity.

"The goal isn’t to standardize writing, it’s to understand and amplify what makes each writer unique."

Q: You’ve worked with emotion detection and language models — how did that shape your thinking when building something for writers?

A: At Microsoft, a lot of my work involved building AI systems that could detect user emotion or intent based on behavior—stuff like subtle patterns in interaction data. It was about helping machines make educated guesses about what someone might be thinking or feeling, given a very specific context.

Story Stream is different, but the underlying challenge is similar: how do you get an AI to understand what matters emotionally in a very human situation?

In fiction, emotions aren’t just random—they’re deliberate, crafted. They often follow recognizable patterns. If a character gets dumped and doesn’t feel sad, the reader knows something else is going on. That dissonance is meaningful. So it’s not just about the words on the page, but the context and intent behind them.

See How Story Stream Makes Revision Smarter and Simpler

Story Stream is designed to help writers see their work more clearly, not just fix typos. If you’re ready to go deeper, check out our editing platform or explore the features that make revision smarter.