A team of former Amazon artificial intelligence veterans has launched Primitive Labs in San Francisco, introducing an AI-driven platform that models human behavior to ensure software products remain customer-centric during the rapid AI development boom. Founded by Rohit Talluri, Jean Farmer, and Gabriel Fong, the pre-revenue startup automates traditional user research by deploying autonomous AI agents that simulate how real people navigate software across desktop and mobile devices.
The Shift from Traditional Focus Groups to AI Simulation
At Amazon, product teams famously maintain an empty chair during meetings to symbolize the customer’s presence and voice. However, as generative AI tools accelerate the pace of software creation, startup founders warn that user experience testing is increasingly sidelined due to tight production deadlines.
Traditional user testing and focus groups often require weeks or months to yield actionable insights. Because of these long lead times, product managers under pressure to ship code frequently skip user validation altogether, risking the release of unintuitive features.
Primitive Labs addresses this bottleneck by replacing manual testing cycles with synthetic behavioral intelligence. The company aims to embed human behavior directly into the software development stack as a foundational building block, enabling engineering teams to observe how specific user demographics react to new features, designs, or marketing decisions before code is deployed.
Deep Roots in Frontier AI and Autonomy
The startup’s leadership brings deep experience from Amazon’s Advanced General Intelligence (AGI) and Amazon Web Services (AWS) infrastructure teams. Chief Executive Officer Rohit Talluri previously worked in Amazon’s AGI Autonomy Lab on computer-use agents and helped launch Nova Act, Amazon’s agentic computer-use model, following the tech giant’s talent acquisition from Adept AI.
Chief Technology Officer Jean Farmer previously led tool-use capabilities for Amazon Nova models and built testing systems for large-scale machine learning infrastructure, including SageMaker HyperPod. Farmer also contributed to benchmarks for the Model Context Protocol (MCP), an emerging open standard connecting AI models to external software tools.
Chief Operating Officer Gabriel Fong rounds out the team with background in product marketing, cloud consultancy, and enterprise account management at AWS and DoiT International. Fong works closely on core product direction and operational execution alongside Talluri and Farmer.
Venture Backing and Early Enterprise Adoption
Primitive Labs recently secured pre-seed funding led by venture firm Andreessen Horowitz through its Speedrun accelerator program. The round attracted a roster of prominent venture funds and angel investors, including Olive Tree Capital, Cloverfield Fund, and Unexpected Investments, alongside former Adept CEO David Luan and industry veterans from OpenAI, Google DeepMind, Meta, Nvidia, and Databricks.
“It’s bringing humans back to the center of a world that’s created by AI,” Talluri stated regarding the startup’s core mission. “That is the goal here.”
The company is currently testing its technology with a select group of early enterprise clients, including private previews with Fortune 500 and Fortune 50 consumer technology and e-commerce brands. The underlying system draws on computational cognitive science and specialized memory architectures designed to reflect human learning patterns, evaluating performance through a metric called behavioral fidelity.
Reshaping the Software Development Lifecycle
The emergence of behavioral simulation signals a potential shift in how digital products are designed, tested, and validated. As automated coding tools produce software faster, synthetic user testing could become standard development infrastructure alongside continuous integration and automated unit testing.
Primitive Labs plans to publish portions of its cognitive research and open-source select components in the coming months. The company is actively recruiting founding engineers and researchers in San Francisco and New York as it joins the a16z Speedrun cohort this month.
Industry analysts will be watching closely as Primitive Labs prepares for its general product availability later this year and plans its next funding round in autumn. The key test will be whether synthetic agents can accurately forecast real-world human behavior, user retention, and digital conversion rates at scale.
