Bern: a clearer purpose for a talking toy

The problem

At Bern, our starting point was broad: could voice AI make a screen-free toy engaging enough for children? As founder and CEO, I had to connect what the technology could support, what children enjoyed, and what parents wanted in their homes. We interviewed parents and tested with children in person.

Give the bear a clear purpose

Parents were uncomfortable with a bear that could talk about anything, even with restrictions. Adding more capabilities would not resolve that concern. Their trust was part of the product problem.

We moved toward language learning, using weekly seasons and daily episodes built around stories and conversation. That gave us a defined experience to build and gave parents a clearer reason to want the bear.

Make the technical problem understandable

The bear took roughly four seconds to respond. I introduced plain-English explanations alongside technical specifications so people outside engineering could understand and question our decisions.

When I explained the audio process to the design team, our main designer suggested breaking the audio into smaller pieces. Engineering thought it could work. I then recruited an outside scientist to advise us and work with the engineers on the hardware application.

The designer proposed the idea; the scientist and engineering team worked through the implementation. Our team tested the approach and reduced response time from four seconds to two. My contribution was creating shared context and bringing the right people into the problem.

What I took away

Bern reached 1,000+ beta testers. The latency test gave us a concrete technical result; it did not establish an improvement in learning or long-term use.

I learned to treat a parent's hesitation as product feedback, even when a child enjoyed the toy. Inside the team, making the problem understandable let someone outside engineering contribute a useful technical idea.