There's a moment I think about often. I was sitting with my grandmother, recording her voice on my phone — not because I had a plan, but because I was afraid of losing her. Not just losing her, but losing the sound of her laugh. The way she paused before telling a story. The warmth in her voice when she said my name.
I realized something in that moment: we spend so much time and money preserving photographs and documents, but the thing we treasure most — someone's voice, their presence, the cadence of their speech — is the hardest to hold onto.
That's when I started wondering. What if there was a tool that did for voices what cameras did for faces? Not just recording, but preserving — with intention, with care, with structure. Something that made it natural and easy for people to tell their stories.
I talked to researchers at UCLA, UNM, NIH, and Northwestern. I learned about cognitive interviews, life review therapy, motivational interviewing, narrative identity — and eventually, the critical importance of memory integrity and safety in AI-mediated conversations. These frameworks have been used in clinical settings for decades to help people recall memories, process experiences, and find meaning in their lives.
What if we combined them? What if we built something that felt like talking to someone who truly listened?
That became Echoes.