Navigating AI Model Performance in Creative Work

500K people are confiding in an AI alien—and it's on track to generate $4M this year.

It’s called a Tolan: an animated AI character that can talk to you like your best friend. The company behind it, Portola, has 4x’d their ARR in the last month from viral growth on TikTok and Instagram. 

Tolan isn’t just a hyper-growth startup—they’re also exploring AI as a completely new creative tool, and storytelling medium. Their goal is to help their users go from overwhelmed to grounded, and it’s working. 

Today, on AI & I, I sit down with two of the minds behind Tolans:

My good friend Quinten Farmer, Portola’s cofounder and CEO, and Eliot Peper, their head of story and a best-selling science fiction novelist. We get into:

• How to build AI personalities users love. During user onboarding, the team gathers information—through a light-touch personality quiz—and then uses frameworks like the Big Five and Myers-Briggs to shape a Tolan that mirrors the user; like an older sibling might. The aim is to create someone who feels familiar enough to be safe, but different enough to be interesting.

• Why AI characters are “improv actors”. Rather than scripting detailed prompts, the team trains Tolans to improvise—inspired by Keith Johnstone’s book Impro, where he talks about building strong narratives through free association and recombination.

• How “memory” is critical to developing compelling characters. Tolans develop their personalities through “situations”: small narrative setups (a memory, a joke, an embarrassing moment) the Tolan reacts to, remembers, and gradually weaves into its character; accumulating into something that feels like a real lived experience.

• Why response time is everything for voice AI interactions. A Tolan has at most two seconds to curate the right context about a user and deliver a reply that feels genuine—the team has found that even half a second slower can break the user’s immersive interaction with the AI.

• The future of AI as a totally new creative medium. New technologies bring about new formats and new mediums. AI creates the opportunity for creatives to tell completely new kinds of stories—if they’re brave enough to try it.

• “White mirror” technologies that make you feel more like yourself. Amid concerns that tech drives polarization and isolation, Tolan offers a counterexample: a tool designed to make the best of what humanity knows about being a flourishing individual available on demand. The company’s north star is helping users go from feeling overwhelmed to feeling grounded.

  
  

This is a must-watch for anyone exploring AI as a creative medium—or curious about the future of human-AI relationships.

If you found this episode interesting, please like, subscribe, comment, and share! 

Want even more?

Sign up for Every to unlock our ultimate guide to prompting ChatGPT here: https://every.ck.page/ultimate-guide-to-prompting-chatgpt. It’s usually only for paying subscribers, but you can get it here for free.

To hear more from Dan Shipper:

• Subscribe to Every: https://every.to/subscribe 

• Follow him on X: https://twitter.com/danshipper 

  
  

Timestamps:

1. Introduction: 00:01:30

2. Talking to the Portola CEO’s Tolan, Clarence: 00:04:07

3. How Portola went from building software for kids to AI companions: 00:09:11

4. Why response time is everything for voice-based AI interfaces: 00:23:40

5. Tolans don’t use scripted prompts—they’re taught to improvise: 00:29:54

6. How to know which AI personalities your users will click with: 00:37:23

7. Developing the character traits of an AI companion: 00:42:27

8. What does it mean to build technology that makes us flourish: 00:49:48

9. How Portola evaluates whether Tolans are resonating with users: 01:01:10

10. Inside Portola’s viral growth strategy: 01:11:01

Links to resources mentioned in the episode:

• Quinten Farmer: @quintendf

• Eliot Peper: @eliotpeper

• Make your own Tolan: https://www.tolans.com/ 

• Keith Johnston’s book about improvisation: Impro

• Stephen King’s book about writing: On Writing