Again in January, the viral sensation OpenClaw kicked off the private AI assistant craze, the place you put in a group of AI brokers in your desktop, give them objectives, and set them unfastened. Quick ahead to Could and now there’s a brand new buzzy pattern that’s able to take off: “proactive” AI (i.e., AI brokers that go to work with out even being requested).
ChatGPT jumped onto the proactive AI bandwagon early, releasing the ChatGPT Pulse characteristic means again in September 2025. Nonetheless in an early preview section and accessible solely to ChatGPT Professional customers on cellular, ChatGPT Pulse delivers a personalised morning briefing primarily based on the subjects you’ve not too long ago broached with the chatbot, together with knowledge out of your linked Gmail inbox and Google Calendar in addition to “direct suggestions” from you.
Now it seems Claude and Gemini are heading down the identical path. Proof discovered within the Claude desktop and cellular apps factors to a possible new characteristic known as “Orbit,” which is a “proactive assistant” that connects to instruments like Gmail, Slack, GitHub, Google Calendar, Drive, and Figma, experiences TestingCatalog.
The unannounced Orbit characteristic, which may very well be unveiled as quickly as Wednesday at Anthropic’s Code with Claude convention in San Francisco, seems to be aimed primarily at builders seeking to handle their manufacturing workflows, permitting Claude to proactively serve up every day briefings that it researched the night time earlier than.
In the meantime, Google seems to be busy at work on a “Proactive Help” characteristic for Gemini, with Android Authority recognizing traces of the characteristic within the Google app.
Based mostly on screenshots unearthed by Android Authority, Proactive Help would piggyback on Gemini’s Private Intelligence performance, taking knowledge out of your linked Google companies (like Gmail, Google Calendar, and Google Drive), your textual content messages and contacts, what’s in your display, and your notifications to ship “customized solutions on the proper time.”
As with Claude’s rumored Orbit characteristic, Proactive Assistant for Gemini has but to be formally confirmed, though Google may unveil it subsequent week at its Google I/O keynote on Could fifteenth.
I’ve reached out to each Anthropic and Google for remark.
All of those “proactive AI” instruments would take at present’s private AI assistants to a brand new degree. Whereas agentic AI instruments like OpenClaw and Claude Cowork depend on outcomes beforehand outlined by the consumer, Pulse, Orbit, and Proactive Assistant seem poised to behave first, sussing out your pursuits and intentions primarily based in your chats, linked apps, and different indicators.
Would these proactive experiences and updates really be helpful? Or just annoying? That each one depends upon the implementation. For instance, I typically discover my Claude Cowork automations to be surprisingly useful (these are automations I requested Claude to arrange), whereas Gemini’s pushy Private Help characteristic (“That may be excellent to your new Manhattan house, Ben!”) has been a significant turnoff.
Proactive AI that easily anticipates your wants with out first needing to be requested is one other piece of the private AI assistant puzzle. Prepare for rising “proactive AI” buzz within the weeks and months forward.
