Startup Stories – Mixergy
Episode #2290: He’s building an AI media empire
Host: Andrew Warner
Guest: Ricardo Vai Santos, Founder of Dream Stories
Date: December 29, 2025
Episode Overview
In this forward-looking episode, Andrew Warner interviews Ricardo Vai Santos, the founder of Dream Stories – a company at the frontier of personalized media using AI to create unique, custom children’s books and looking ahead to fully individualized movies and media. Ricardo unpacks the evolution from personal needs, the technical challenges, business model lessons, and a bold vision for the future of generative AI not just as a mass content creator, but an enabler of deeply personal “bespoke” media experiences.
Key Discussion Points & Insights
The Genesis of Dream Stories: From Personal Motivation to Massive Scale
- Personal Experience Sparked the Idea: Ricardo’s background in media (Spotify, Kinko) combined with the desire to create adventurous, meaningful content for his own children was the seed.
- “There was a disparity between the content that he was reading and the things I remembered from childhood...I wanted him to be Tintin,” (06:04)
- Early Inspiration: Witnessed the popularity of personalized songs (even before AI) on Spotify, which led to the “what if” of scalable, personal, AI-driven content.
- “If you give people the tools to creation, they will do things that are not necessarily an extrapolation of what we do today.” (04:38)
Technical Hurdles in Personalized AI Storytelling
- Difficulty with Image Consistency: Unlike generating a single image, keeping characters “on-model” throughout a multi-image book was a unique challenge.
- “The challenge with the AI…is the consistency of images, actually kind of complex problem.” (08:14)
- “If it’s already hard to keep consistent with one character, it becomes exponentially harder with multiple.” (26:17)
- Real-World Users vs. Tech-Savvy Users: Technical friction for end-users (e.g., poor photo uploads) forced Dream Stories to create robust processes for mass market use, not just tech demos.
- “The average consumer is not looking for many tries, they’re looking for something that kind of works.” (09:04)
- Use of Synthetic Data: Dream Stories found users re-uploading AI-generated images (because they were better), which led to innovations in how synthetic images could be used to improve model training.
- “I realized this is what people in AI would call synthetic data… It actually works really well to increase the quality.” (51:02)
Evolution from One-Off Product to Service Mindset
- Recurring Rituals Drive Business: The nightly bedtime reading ritual offered repeat engagement – a platform for building ongoing episodic content.
- “Every night...you can actually build something that is not a product, something that’s a service.” (21:01)
- Beyond Get-Rich-Quick: Ricardo avoids focusing on revenue alone, instead prioritizing the long-term vision (though mentions $3 million as an outdated annual figure).
- “We’re not talking about revenue anymore… what the revenue we’ve had is kind of in support of the larger vision.” (01:21)
Distribution and Business Model Insights
- Distribution is Harder than Technology: The biggest barrier to copying Dream Stories isn’t the tech—it’s efficient distribution and profitable paid acquisition.
- “Execution is not so much the problem as the distribution part...getting people to actually use it.” (28:25)
- “Usually what I do is paid acquisition…What’s difficult is to actually get the economics to work for you, right? Particularly because it’s an auction model.” (29:06)
- Focus on Customer Retention & Repeat Purchases: Rather than novelty, Dream Stories seeks recurring customer value (multiple books, episodic stories)—not a one-time personalized object.
- “The category exists…But what you then need to do is to convert that novelty purchase into a repeat customer that buys many things.” (32:14)
- Subscription Philosophy: He discourages confusing “subscribe & save” discounts; true success comes from building a product people naturally want to subscribe to.
- “You want to build like a service…in which subscription is just a feature of that service.” (33:30)
Vision for AI-Personalized Media
- The Future is Bespoke Content: Mass broadcast media will coexist with highly individualized content streams, made affordable and possible as AI generation costs fall.
- “I do think there would continue to be movies… but also think that there’s going to be this emergent behavior of creating content just for your own consumption.” (15:30)
- Dream Stories as a Platform: Books are just the beginning; the vision is customized video, music, coaching, and beyond.
- “We want to go into other media right…video is still quite expensive, but it will come.” (23:48)
- Opportunity for Others: Encourages entrepreneurs to look for “personalization opportunities”—tools and services, even in coaching or business settings, that use AI to customize experience around the end-user’s context.
- “If I would have a fund, would literally be that—personalized VC…essentially chase personalization opportunities.” (42:19)
Handling Customer Support at Scale with AI
- AI-Powered Support with Human-Like Persona: Uses Intercom’s Fin AI agent but stresses the need for heavy tailoring and persona definition.
- “It takes a lot of tailoring to get it to where you want it to be…I always crafted like a specific Persona in that I want my agents to impersonate, whether humans or AI.” (55:22)
- Stays Close to Customers: Ricardo talks to customers daily, generates product ideas from their hacks, and watches for emerging usage patterns.
- “Primarily, I built this product for myself, so I think it’s straightforward…But when you’re doing a product for families…there’s always going to be people that encounter issues.” (48:36)
Memorable Quotes & Notable Moments
The Vision of Personalized Media
“If you give people the tools to creation, they will do things that are not necessarily an extrapolation of what we do today.”
— Ricardo (04:38)
“There’s going to be this emergent behavior of creating content just for your own consumption, such as taking photos for yourself...It’s not so foreign when I tell you in that way.”
— Ricardo (15:30)
Technical Challenge with AI Consistency
“The consistency of images is actually kind of complex problem...The average consumer is not looking for many tries, they’re looking for something that kind of works.”
— Ricardo (08:14, 09:04)
On Why Personalized Kids’ Books Work
“The one time of the day I think is kind of sacred for families tends to be bedtime...the parents tend to be very restrictive about what they give their kids.”
— Ricardo (19:27)
Why Distribution Beats Product
“Execution is not so much the problem as the distribution part...getting people to actually use it.”
— Ricardo (28:25)
On Subscription Products
“When you mix [one-time purchase and discount subscription], you’re just adding confusion...People are not dumb. They will think there’s a catch.”
— Ricardo (36:58)
User-Generated Synthetic Data Discovery
“I realized this is what people in AI would call synthetic data...it actually works really well to increase the quality.”
— Ricardo (51:02)
Key Timestamps
- 00:37 — How many books and characters Dream Stories has produced
- 01:21 — Shifts focus away from revenue to vision
- 02:30 — How childhood reading routines inspired the idea
- 04:38 — Early thinking on future of personalized media
- 08:14 — The technical problem of image and character consistency in AI books
- 12:29 — Early collaboration with co-founder and building physical product
- 15:30 — Comparison to the iPhone’s effect on photography; rise of highly personal media
- 19:27 — Why bedtime is strategic for recurring SaaS-like product
- 23:48 — Vision to expand into personalized video and other media (cost constraints discussion)
- 28:25 — Real hardest part: Going from product to mass distribution
- 32:14 — Differentiating on episodic, repeatable user engagement
- 33:30 — Thoughts on subscription models and avoiding user confusion
- 42:19 — Blue-sky on “personalization opportunities” for other entrepreneurs
- 48:36 — Talking to customers every day; product insights from customer hacks
- 51:02 — Discovery of using synthetic AI-generated data for better model results
- 55:22 — AI-powered customer support and the importance of defining the right agent persona
Final Thoughts
For aspiring entrepreneurs:
- Focus less on technical novelties and more on solving for repeat engagement and distribution economics.
- Personalization is just getting started—the next wave is in products and services deeply tailored to individual or family context, whether emotional, educational, or even business-related.
Ricardo’s core lesson:
“The problem is not so much execution as much as the distribution part…But what I can tell is the way I go about it is figure out a model in which you retain customers.”
(28:25)
Big takeaways:
- Personalized media is an open field with far-reaching consequences—both technically and behaviorally.
- There’s vast room for new business models, products, and even coaching/enterprise applications around hyper-personal AI experiences.
- The truly “hard” part for founders is rarely the technology; it’s creating sustainable, repeatable value, and earning a loyal, returning customer.
For More
- Dream Stories – [Link to company]
- Ricardo Vai Santos on personalized AI media
- Mixergy’s archive for more founder interviews
Podcast ad spots, intro/outro, and midrolls were skipped for focus on content.
