Don and Jean were in the same Math and Science Academy in high school. They reconnected as Jean was building Primitive Labs. After we met Ro, we knew this was a company we wanted to back from the earliest stages.
Ro, Jean, and Gabe are building AI agents that simulate how people use software. Product teams can use those simulations to identify where customers get confused, hesitate, or leave before shipping a change.
The Problem We Know Firsthand
At Nolla, Don and Sean have seen how small UI changes affect conversion. Understanding where people drop off means reviewing their behavior, forming a hypothesis, and testing another version. We discussed that process in our first conversations with Primitive.
What Primitive Labs Does
Primitive uses customer behavior data to model how different people navigate a product. Its agents move through the interface visually and account for differences in attention, patience, and motivation. A customer might miss a button, misunderstand a price, or abandon a checkout. The team is building simulations that account for those decisions.
Product teams can use this when testing a checkout, revising onboarding, or changing a subscription flow. Simulating those experiences gives them another way to evaluate a change before releasing it to customers.
The Team
Ro, Jean, and Gabe worked together at Amazon before starting Primitive Labs. Ro and Jean worked on machine learning infrastructure and later at Amazon AGI. Gabe focuses on the platform and product. All three write code and speak with customers.
In our conversations, Ro and Jean spent time on validation: which customer data to use, how to benchmark the simulations, and how to establish whether the results reflect human behavior. Those questions determine whether a product team can use the output to make a decision. Their approach to that work helped us decide to invest.
See what they're building at primitivelabs.ai.
Marble Creek Ventures is an investor in Primitive Labs.
