AI Agent Startup Instinct Eyes $10 Billion Valuation as Consumer Adoption Stirs Privacy Concerns

The market for consumer artificial intelligence agents is accelerating as developers shift away from standalone chatbot interfaces toward tools embedded directly in messaging workflows. Among the platforms drawing early attention in Silicon Valley is Instinct, an invite-only AI assistant that connects with users over iMessage and WhatsApp, alongside Meta's newly released Muse assistant, according to a report by Wired (https://www.wired.com/story/i-finally-found-an-ai-agent-worth-the-risk/).

Instinct, which launched in private beta in February, integrates directly with email, calendar, and messaging accounts to proactively suggest and execute administrative tasks. The startup is currently in talks to raise $1 billion in fresh funding on top of $350 million already raised, which would value the company at $10 billion, according to reporting by The Information cited by Wired.

The sudden interest in messaging-first agents coincides with rapid early downloads for Meta's Muse, which launched earlier this month. Third-party estimates indicate Muse has surpassed 900,000 downloads, ranking as the top free application on Apple's App Store. However, the rollout has exposed early technical compromises. Ars Technica reported that Muse initially shipped with a security vulnerability that could allow attackers to execute arbitrary commands on a victim's Mac, while reporting from 404 Media found that some phone reservation calls made by the assistant were handled by human call center workers rather than automated software.

These tools reflect an ongoing product shift in how users interact with autonomous AI. Early personal productivity experiments like Claude Cowork functioned largely as standard chat interfaces requiring constant user prompting. Other competitors, such as Lindy, have taken a more aggressive stance, sending daily morning text messages and joining video calls uninvited. By contrast, Instinct focuses on suggestions delivered within existing messaging threads to handle specific administrative chores.

When applied to personal workflows, consumer agents have shown tangible utility alongside acute operational flaws. In one test reported by Wired, Instinct identified a 90-minute schedule change by Alaska Airlines that qualified a traveler for a full refund on a non-refundable ticket, successfully claiming a $550 reimbursement and rebooking the return itinerary.

Yet the software introduces significant data handling and operational liabilities. Instinct's terms of service allow the company to use customer conversations to train its AI models. Users and early testers have reported that disconnecting the service from an email account did not purge retained inbox copies. The assistant has also triggered platform rate limits; one venture capitalist was banned from dining service Resy after Instinct pinged its API roughly 200 times in an hour. In another instance, the bot canceled a delayed food delivery order without securing the requested refund, forfeiting $64.

The operational friction underscores the gap between backend capabilities and consumer execution. Technology journalist Jasmine Sun noted that "most people's problems are not software-shaped, and most won't notice even when they are." As technology executives cite agentic automation to justify major capital spending on data center infrastructure, user tolerance for privacy trade-offs and automated errors will test the commercial viability of personal AI. Instinct did not respond to requests for comment from Wired.