For tens of thousands of years, storytelling has been ubiquitous in society, across all countries and cultures. But every day, countless stories are lost or forgotten for many reasons, from family members passing away to employees leaving companies.
Brayan Romero (HES '22, SEAS visiting student '22) understood the importance of preserving meaningful stories long before attending Harvard. Within a few years, he faced the heartbreaking losses of both his mother and aunt. During this time, he brought his cousins together virtually to reflect on their lives and the lessons they left behind.
“When somebody passes away, their untold stories and wisdom are lost forever,” reflected Romero. Sharing these stories and learning the wisdom that came with those stories became “the most meaningful way to honor loved ones.”
The idea of preserving stories stayed with Romero for years. In 2023, as generative AI tools like ChatGPT and DALL-E gained popularity, Romero began exploring how people could use GenAI to positively impact their families, cultures, and businesses. He and a few friends developed the first prototype in a week. Less than a year later, Romero and Onchanok (Kamsai) Nawapruek (GSD ’24) transformed this idea into a thriving commercial venture: Storyfave.
During a recent interview with the Harvard Innovation Labs, the co-founders shared how they built Storyfave as well as their plans for growth in 2025 and beyond.
A serendipitous meeting at the i-lab
In fall 2023, Nawapruek visited Harvard Business School for the Kevin O’Leary Shark Tank event, and later stopped by the Harvard i-lab to “check out the space.” While standing in the i-lab kitchen, a mutual friend introduced Nawapruek and Romero, and they started talking about what Romero was building.
Nawapruek shared her background in design and education, and asked how she could help. They agreed to meet the following week to discuss Storyfave. Initially scheduled for an hour, they ended up talking for more than three.
“I was so passionate to work on this, because I related to the idea so much,” said Nawapruek.
Since meeting by chance at the i-lab, Romero and Nawapruek have “talked to every advisor,” and Nawapruek added that they are very grateful for the advisors and mentors that decided to work with Storyfave, who they found through the i-lab Mentor Match program.
Solving AI challenges in product development
As they built an AI agent platform to immortalize stories, Romero and Nawapruek agreed that speed and quality were critical to the user experience and their overall vision — but, for the reader, meaning mattered most.
That’s why we create meaningful books with at least one human in the loop — not just another AI-generated book. Humans are in the driver’s seat, not AI.
Brayan Romero
One of the greatest challenges of achieving this quality, according to Romero, was addressing a well-known challenge in the field of computer vision, which IBM defines as “teaching computers and systems to derive meaningful information from digital images, videos, and other digital inputs.” Romero shared that they’ve focused on training their models to address what he describes as “one of the Holy Grails of computer vision – premium-quality multi-character consistency,” or ensuring that the characters in the story are recognizable on every page.
Addressing this challenge requires Storyfave to ask its customers for images of the characters they want to include in their story. These images are used to privately train a custom model for each user, ensuring that the data remains secure and is not used to train subsequent models. The result is highly personalized stories featuring characters that closely resemble the real people being depicted. Beyond character consistency, Storyfave also prioritizes scene consistency, audio narrations, physical book options, and more — all aimed at making the final product more meaningful and effective.
Prototyping a physical book
Storyfave didn’t just want to be an AI platform for creating digital storybooks. From the beginning, Romero and Nawapruek were passionate about printing beautifully designed, heirloom-quality books that customers could cherish for years.
To show potential book printers what they wanted, Romero and Nawapruek used the Storyfave platform to create their first story, printed each individual page, purchased supplies from a BLICK Art Materials store, and “glued together” the book themselves.
"I had a similar story to Brayan, in that my grandpa passed away as I was applying to Harvard GSD,” shared Nawapruek. “He didn’t live long enough to know I got admitted. That was something that stayed with me for a while. So our first story was about him and the lessons he taught me.”
With this sample in hand, they began contacting vendors to see if they could print a book that had a similar look and feel. After contacting more than 100 vendors, they finally found the right fit.
Maintaining a human touch
Storyfave secured their first customers soon after identifying a print vendor. As they’ve worked with customers on telling their own stories, they’ve continued to emphasize the importance of manual edits and refinements to images and text.
Once customers answer Storyfave’s questions that are used to generate a story, and submit the necessary photos of the characters, the customer receives a first draft of the story. The customer then may edit the text manually and select the images they like the most. Storyfave’s team is also available to help customers make revisions. In this process, the human is the director of the book, not the AI.
Romero and Nawapruek shared that Storyfave “retrains and fine-tunes its model” based on customer edits to improve the quality of the book, ensuring that future drafts are even better. They emphasized that any retraining or fine-tuning is done exclusively for that specific book, and the data, including images, is kept secure and never used for other books or shared with other users.
Expanding to corporate storytelling
In less than a year, Storyfave has printed more than 800 books for dozens of customers on topics ranging from a fictional children’s tale about magical camels to nonfiction stories about loved ones’ lives. As Storyfave continued to refine their platform throughout 2024, Romero and Nawapruek said they have reduced the time it takes to make a print-ready story from “weeks to a few hours — and ultimately it will be minutes.”
Looking toward the future, Storyfave sees a significant opportunity in working with businesses to engage their customers, partners, and investors through storytelling. The company recently introduced an enterprise offering using AI agents for organizations to work with Storyfave on stories about their mission, founder journeys, customer success, and more.
“To help businesses, our algorithm can quantify emotional connection and demonstrate how our product can improve sales. This aligns with research published in the Harvard Business Review, which describes the Emotional Connection Score as a significant factor in driving customer repurchases and enhancing brand differentiation,” said Romero.
While the use cases for Storyfave will continue to expand, Nawapruek and Romero emphasized that the company’s commitment to helping humanity create their “most meaningful stories” is expected to remain unchanged.
