
What happened
Noeri Inc. launched with $110 million in funding led by Menlo Ventures, with Madrona Ventures, Kindred Ventures, Wing Venture Capital, NEA, Snowflake and Cisco Ventures also participating.
Why it matters
Noeri's software learns a company's workflows and proposes tweaks — different instructions, tools or models — each explained and sent to a human for approval, aiming to offer an alternative to paying for top-tier models or settling for weaker ones.
WHO IT HITSEnterprise IT and operations teams weighing AI model costs may get a tool that recommends workflow changes for human sign-off while keeping ownership of their systems.
Summaries like this, every morning.
Noeri's pitch rests on two research threads from its founders. Guestrin pioneered a self-distillation technique that lets AI models become their own teachers, turning mistakes into lessons. Choi's work on agentic orchestration showed that smaller models coordinating teams of AI agents and tools can outperform much larger frontier models at lower costs.
The idea came from helping a Stanford microbiology and immunology team use OpenAI's models more cost-effectively. Choi said she and Guestrin built a small AI model that learned to use OpenAI's older models and, surprisingly, fine-tuned it to perform better than GPT-5 on that team's projects.
Noeri's name means 'intelligence, in the plural,' reflecting the founders' belief that AI won't be dominated by one or a few powerful models. The startup says it is working with partners across multiple industries, including some investors chosen to expose its systems to a wide variety of work. Menlo Ventures Partner Venky Ganesan said the firm moved quickly to lead the round because the founders are putting world-class research to work inside companies at the same time.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.
Alibaba launched Qwen4-Max, built on a new Qwen4 architecture, with a 2.4 trillion parameter backbone that act…
Qualcomm CEO Cristiano Amon told Yahoo Finance on November 12 there is "no AI correction in Qualcomm," pointin…

A Qiita tutorial from engineer skill-sharing platform DokuPro walks through collecting thumbs-up/thumbs-down r…

Altimeter Capital's Brad Gerstner told CNBC he is placing a $5 billion order in Anthropic's IPO, five times hi…

Corma, an AI cybersecurity startup that came out of stealth in August, raised a $125 million Series A led by N…

Duco moved its agentic AI stack to Amazon Bedrock and AgentCore, reducing main context from 200,000 tokens to…
