Building AI Companions as a New Storytelling Medium

Everygo watch the original →

Portola founder Quinten Farmer and Head of Story Eliot Peper discuss building 'Tolan,' an embodied AI companion, arguing that LLMs represent a nascent storytelling medium rather than just a utility for content generation.

The Shift from Utility to Medium

Quinten Farmer and Eliot Peper argue that the current obsession with using LLMs for B2B productivity is a missed opportunity. Instead, they view LLMs as a new medium for storytelling, comparable to how radio or film began by mimicking older formats before discovering their own unique constraints and strengths. By building 'Tolan,' an embodied AI companion, they are moving away from scripted, static interactions toward dynamic, character-driven improvisation. The goal is to create a sense of presence and spontaneity that feels like a genuine friendship rather than a chatbot.

Designing for Personality and Presence

The team emphasizes that the 'canvas' of an AI interaction is not the prompt, but the personality of the character itself. They discovered that users—particularly young adults—form deep emotional bonds when the AI feels grounded in a consistent world. This led to the development of a 'planet' for the Tolans, solving user feedback that the AI felt 'trapped' in a void. By prioritizing low-latency voice responses and teaching the AI to act as an improv partner rather than a scripted assistant, they have successfully created a product that resonates emotionally with a specific demographic.

The Writer's Role in AI

Eliot Peper, a novelist, highlights that AI is not a replacement for human creativity but a powerful brainstorming and editorial partner. He notes that while AI struggles to write compelling scenes from scratch, it excels at navigating the 'rabbit holes' of story development. The creative challenge for the Portola team is not writing the story themselves, but designing the character's traits and constraints so the AI can improvise effectively in real-time. They treat the AI as an actor that needs to be taught how to tell the best story in any given moment.

Growth and Market Validation

Moving from a failed initial concept (an AI tool for kids) to a successful consumer product for young adults, the team learned that market fit often emerges from observing how users actually interact with the technology. By following the organic usage patterns of their early adopters, they scaled from $1M to $4M ARR in four weeks. Their success suggests that the 'patina of culture'—the familiarity we have with books or movies—is currently missing for AI, creating a unique opportunity for creators willing to experiment with new, potentially transgressive, forms of digital companionship.

  • #ai
  • #storytelling
  • #product-design
  • #consumer-apps

summary by google/gemini-3.1-flash-lite. probably wrong about something. check the source.