Our Story
Support that resolves
We believe that setting up Duckie should be just as seamless as your customers' experience interacting with Duckie agents.
The problem
of tickets are repetitive
avg. first response time
agent time on actual issues
of customers expect instant help
Most importantly, the process to actually set up your AI agent team is way too complex and takes months to show real value.
How we build
AI should take actions
We don't build chatbots that explain how to do things. We build AI that actually does them, so you don't have to keep closing tickets manually.
AI should save money
Implementing AI solutions should save your company money without compromising on quality. Not add more bloat.
And be easy to set up
Your support lead should be able to set up your AI agents using plain English without engineering help or a long timeline.
The origin
From on-call frustration to breakthrough
It started with a simple problem: Valerie was tired of answering the same questions during on-call rotations. So she built a bot—something that could search through documentation, understand context, and actually help.
That internal tool worked so well that teammates started asking for access. Then other teams. Then other companies.
The pattern was clear: everyone was drowning in repetitive support work, and AI had finally gotten good enough to help.
We didn't want to build another chatbot that just answers questions. We wanted to build AI that actually resolves issues.
Built by engineers who lived the problem
Valerie
Co-founder & CEOFormer senior engineer at LinkedIn. Built the first version of Duckie to solve her own on-call pain.
Joel
Co-founder & CTOFormer senior engineer at Netflix and C3 AI. Obsessed with building AI that's actually reliable.
