Industry insiders say the next big thing for AI is “proactive” systems, or agents that can anticipate and meet a user’s needs before the user even knows they have a need.
One startup looking to make headway in this space is IrisGo. The company, which closed a $2.8 million seed round led by Andrew Ng’s AI Fund earlier this year, is building a desktop companion for PCs that can learn your daily workflows and automate them with limited or no human prompting.
Iris was co-founded by Jeffrey Lai, a former Apple engineer who helped build the Chinese version of the company’s automated assistant, Siri. (It’s a little sneaky, but Iris is Siri spelled backwards.)
The central idea is simple. Once you show Iris how to do something, it remembers that process for future automatic use. No repeated instructions required.
In a conversation with TechCrunch, Lai ran a demo showing how Iris could learn how to order coffee online. As I watched, Iris was recording the process of picking a latte at Philz Coffee (a popular Bay Area chain), entering her credit card information, and pressing the purchase button. Rai then asked Iris to repeat the command herself. The agent dutifully obeyed.
Of course, buying coffee is not the real purpose. Instead, the system is expected to automate the entire business-related tasks. Iris comes with a built-in library of “skills” such as email drafting, invoice processing, report writing, document summarization, and many other ready-to-use automated workflows. At the same time, Iris learns from your desktop behavior and automatically adds those tasks to your list of potential action items.
The application also includes a coding assistant (similar in concept to OpenAI’s Codex and Anthropic’s Claude Code) designed to help developers.
“Our target audience is knowledge workers, white-collar companies. There are a lot of repetitive tasks that they do every day,” Lai said, noting that despite the sheer power of today’s frontier models, AI-assisted office work can still feel incredibly manual and repetitive. The goal, he says, is to move away from that and toward a more fully autonomous workflow, where humans work on high-level conceptual work and agent systems handle all the paperwork in the background.
A particularly attractive feature of Iris is that it is designed to process large amounts of data on-device, offering stronger privacy protections than other applications that rely heavily on the cloud. Lai said the system is still a hybrid architecture. That means larger, more complex tasks will ultimately be processed through the cloud, but the company promises that cloud processing “will only occur when explicitly authorized by the user and with end-to-end encryption.”
One of Iris’ strategies for expanding was to gain credibility through engagement with prominent people and organizations. The support of Ng (particularly co-founder of the formative deep learning research team Google Brain) was helpful. Through their mutual connections, Lai was able to set up a meeting with Ng. Both are graduates of Carnegie Mellon University. Lai and his co-founders demoed Iris at the conference, and Ng’s AI Fund ultimately led the startup’s seed round. Nvidia and Google are also backing the company.
IrisGo recently released beta versions of its macOS and Windows apps. The company is also currently working on deals with laptop companies to pre-install the app on new devices. The company recently signed a similar deal with Acer, and Lai said he expects the company to be able to sign similar deals with other equipment makers soon.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
