When Quilty burst onto the scene earlier this year, the AI startup made a bold promise – its tool could accurately predict a film’s box office success just by analyzing the script. But when industry insiders actually put the product through its paces, reality hit hard. The platform predicted the script for Christy, which became a box office disaster, would outperform Sinners, the film that went on to win Oscars and dominate theaters. It’s another reminder that AI’s promises often outpace its actual performance, especially in creative industries where human intuition still reigns supreme.
Quilty made waves when it hit the trades earlier this year with a proposition that sounded almost too good to be true – feed it a screenplay, and it’ll tell you whether you’ve got a hit or a miss on your hands. For an industry that’s famously unpredictable, where nobody knows anything and every studio executive lives in fear of greenlighting the next Cutthroat Island, this seemed like the holy grail. An AI system that could cut through the gut feelings and give you cold, hard data on what works and what doesn’t.
But then people actually started using it. According to testing documented by The Wrap, Quilty’s predictions didn’t just miss the mark – they got things spectacularly wrong. The platform analyzed the script for Christy, Sydney Sweeney’s boxing drama that The Guardian reported became one of 2025’s biggest flops, and predicted it would outperform Sinners. That’s the same Sinners that went on to become an Oscar-winning blockbuster, raking in hundreds of millions and dominating awards season.
It’s a humbling moment for a company that positioned itself as having cracked the code on predicting audience taste. The entertainment industry has always been resistant to formulaic approaches – William Goldman’s famous line that “nobody knows anything” has been the unofficial motto of Hollywood for decades. Quilty’s founders thought they could change that equation with machine learning and data analysis, but the results suggest that whatever patterns the AI found in successful scripts, they don’t translate to predicting real-world performance.
The company’s pitch followed a familiar pattern in the AI world. Like many startups in the space, Quilty’s founders talked about “democratizing” the film industry by giving up-and-coming creatives access to tools that could level the playing field. It’s a compelling narrative – instead of needing connections to studio executives or established agents, any aspiring screenwriter could get objective feedback on their script’s commercial potential. The promise was that data would replace gatekeeping.
But the Christy versus Sinners debacle exposes the fundamental challenge of applying AI to creative work. Film success isn’t just about story structure, character arcs, or dialogue quality – all things that could theoretically be analyzed by algorithms. It’s about timing, casting, marketing, cultural moments, and dozens of other variables that don’t show up on the page. A script is just a blueprint. The final product depends on thousands of decisions made during production, post-production, and distribution.
The entertainment industry has been grappling with AI’s role for the past few years, from writers’ strikes over AI-generated content to concerns about synthetic actors. But tools like Quilty represent a different kind of AI intrusion – not replacing human creativity, but claiming to evaluate it more accurately than humans can. That’s arguably more threatening to the people who’ve built careers on their ability to spot talent and predict trends.
Industry veterans remain deeply skeptical. If an AI can’t tell the difference between a script that’ll bomb and one that’ll sweep the Oscars, what value is it really providing? The gap between what the technology can actually do and what companies claim it can do keeps widening. In an industry built on storytelling, it turns out the biggest fiction might be AI’s own origin story about its capabilities.
Quilty’s stumble comes at a moment when AI companies across sectors are facing increased scrutiny over whether their products deliver on their promises. From chatbots that hallucinate facts to self-driving features that require constant human supervision, the pattern of overpromising and underdelivering has become impossible to ignore. The entertainment vertical is just another arena where the technology isn’t quite ready for prime time.
What makes this particularly interesting is that Hollywood should theoretically be ideal territory for AI predictions. There’s decades of box office data, script archives, audience demographics, and critical reviews to train models on. If machine learning can’t crack this problem with all that information available, it raises questions about where else the technology might be hitting similar walls.
For screenwriters and filmmakers, Quilty’s failure might actually be reassuring. It suggests that despite all the algorithmic analysis and data-driven decision-making creeping into every industry, there’s still something fundamentally human about stories that resonate. The magic that made Sinners connect with audiences while Christy fell flat can’t be reduced to patterns in a dataset.
Quilty’s miscalculation on Christy versus Sinners isn’t just an embarrassing miss for one startup – it’s a case study in the broader challenge of applying AI to creative industries. While algorithms can crunch data and identify patterns, they struggle with the human elements that actually determine whether a film connects with audiences. As entertainment companies continue exploring AI tools, Quilty’s example serves as a reminder that technology can inform decisions but can’t replace the instinct, timing, and cultural awareness that separate hits from flops. For now, Hollywood’s old wisdom holds true – nobody knows anything, and apparently, neither does the AI.
