IrisGo, a startup backed by $2.8 million in seed funding led by Andrew Ng's AI Fund, is developing a proactive desktop companion for personal computers. Co-founded by Jeffrey Lai, a former Apple engineer who contributed to the Chinese version of Siri, the company aims to build agents that can anticipate and execute user tasks without constant prompting. The core philosophy relies on the system learning a workflow after a single demonstration, automating the process for future use. In a recent demonstration, Lai showed Iris automatically ordering coffee online by recalling the steps taken to select a drink, input credit card details, and finalize the purchase. The application is designed primarily for knowledge workers in white-collar environments to reduce repetitive administrative tasks. It includes a built-in library of pre-made skills such as email drafting, invoice processing, report building, and document summarization. Additionally, the software observes desktop behavior to identify potential automation opportunities and integrates a coding assistant to help developers. A distinguishing feature of IrisGo is its focus on data privacy through on-device processing. While the architecture is hybrid, allowing complex tasks to be handled via the cloud, the company states that cloud processing only occurs with explicit user authorization and utilizes end-to-end encryption. This approach aims to provide stronger privacy protections compared to cloud-heavy competitors. Lai achieved the funding and strategic partnerships through high-profile networking. He connected with Ng, a co-founder of Google Brain and Carnegie Mellon University alumnus, through a mutual contact. During a demo, Ng's AI Fund agreed to lead the investment round. Nvidia and Google also provided financial support for the startup. The company has recently released beta versions of its applications for both macOS and Windows. IrisGo is currently negotiating deals with laptop manufacturers to preinstall the software on new devices. It has already secured a partnership with Acer, with plans to expand to other hardware makers soon. The startup seeks to shift the workflow paradigm from manual clerical work to autonomous background automation, allowing humans to focus on high-level conceptual tasks.