Yutori, which is building personal AI assistants that automate everyday digital tasks, has emerged from stealth with $15m in seed funding.
Yutori is pioneering an agent-first approach that unlocks accurate and reliable personal AI assistants capable of executing tasks across the web.
“Productivity isn’t about cramming more into your day — it’s about reclaiming your attention for what truly matters, and amplifying the outcomes of the time you give something. Yutori’s mission is to build the best AI assistants to make space for the meaningful things in life,” said Devi Parikh, co-CEO and co-founder of Yutori.
Today’s frontier AI models enable great chatbots but can’t complete tasks autonomously. Yutori was founded to build personal AI assistants that are so reliable that they become the primary drivers of action on the web. These assistants will handle tasks and actions ranging from scheduling and communications to authentication and transactions, effectively coordinating users’ lives — creating a world where everyone has access to a team of digital assistants.
This shift will completely redefine how people interact with the web, but to get there, an entirely new approach is required. Current agentic applications built around LLMs propagate errors over long sequences of actions involved in complex tasks, leading to degraded accuracies and as a result, a poor user experience. To make LLMs effective at completing tasks, Yutori is developing an agent-first approach – rather than an LLM-first approach – that unlocks performance across planning and execution.
“When you’re reimagining, from scratch, the interface between consumers and the web — you can’t be focused just on the AI models or just on the orchestration or application layers,” said Abhishek Das, co-CEO and co-founder.
“To raise both the floor and ceiling of human productivity on the web, you need to innovate across the stack in a coherent and tightly coupled way. That’s what we’re doing at Yutori.”
Yutori’s approach focuses on post-training foundation models to be agentic, implementing a multi-agent system that can execute multiple tasks and sub-tasks in parallel, unlocking superhuman performance. The company is pushing the frontier of various post-training techniques including reinforcement learning, test-time search, and model-in-the-loop flywheels atop open-source models with commercial rights. Combined with tightly integrated generative product interfaces, these components enable Yutori to deliver more accurate and reliable AI assistants that consumers can trust.
“What differentiates chatbots from agents is an external environment. The web is the ultimate digital environment — if a task can be done digitally, it can be done via the web. But the web is dynamic, non-deterministic, and noisy; which means mistakes are inevitable and the key agentic skill is resilience,” said Dhruv Batra, chief scientist and co-founder at Yutori.
“Yutori’s agent-first approach will unlock superhuman performance on this grand challenge.”
Yutori was founded by AI researchers and friends Parikh, Dhruv Batra, and Das, who have nearly 50 years of combined experience in AI research. Parikh was most recently a senior director at Meta, leading multimodal generative AI research, supporting Llama 3, Make-A-Scene, Make-A-Video, and Emu, among other initiatives. Batra was a senior director at Meta, leading research in embodied agents that supported the multimodal assistant in the Ray-Ban Meta glasses and FAIR’s Boston Dynamics’ demos. Das’ PhD thesis, Building Agents that can See, Talk, and Act, was recognized by a AAAI / ACM SIGAI doctoral dissertation runner-up award in 2020. The rest of the founding team includes the multimodal post-training tech leads of Llama 3 and Llama 4, as well as engineers and builders from Meta, Gemini, Minion, CMU, and Stanford.
The funding was led by Radical Ventures, with participation from Felicis and over a dozen stellar angel investors, including Fei-Fei Li and Jeff Dean, as well as investors Elad Gil, Sarah Guo (Conviction) and Sandhya Venkatachalam (Axiom). Technology leaders such as Amjad Masad (Replit), Akshay Kothari (Notion), Oliver Cameron (Odyssey), Guillermo Rauch (Vercel), Soleio, and Logan Kilpatrick (Google AI Studio), among others, participated in the round.
Yutori’s seed funding will be used to accelerate the development of its agent-first approach, grow its engineering and design teams, and prepare for its product launch. This spring, early adopters will have the chance to participate in a closed beta of Yutori’s first product offerings.
Jim Cornall is editor of Deeptech Digest and publisher at Ayr Coastal Media. He is an award-winning writer, editor, photographer, broadcaster, designer and author. Contact Jim here.
