Instinct entered a private beta in February and quickly attracted attention from investors. The startup has already secured $350 million and is reportedly courting an additional $1 billion, which would lift its valuation to roughly $10 billion, according to The Information. The service integrates with a user’s email, calendar and popular messaging platforms, allowing commands to be sent via iMessage or WhatsApp.
Early adopters note that the agent’s “form factor”, a conversational interface inside familiar messaging apps, eliminates the need for a separate web portal. Unlike many chat-based assistants that rely on open text boxes, Instinct pushes suggested actions through short prompts, then executes them on the user’s behalf. Reviewers have praised this approach as more natural than the noisy notifications of competing tools such as Lindy.
The author’s first experiment involved a work trip to Venice. By feeding the itinerary into Instinct, the agent secured restaurant bookings, including several off-the-beaten-path venues suggested via WhatsApp. The seamless reservation process appears to be a common entry point for new users, echoing reports that other agents sometimes rely on human operators for phone calls, as noted by 404 Media.
A more demanding test came when a scheduled appearance at the WIRED World Fair conflicted with a pre-booked New York trip. Instinct detected a schedule shift by Alaska Airlines, identified eligibility for a full refund, canceled the original ticket and rebooked a one-way return to San Francisco, saving the author about $550. The outcome illustrates the potential for AI assistants to handle complex travel logistics without direct user intervention.
Security concerns quickly surfaced. Users have reported that the agent retained copies of their inboxes even after disconnection, and a venture capitalist was temporarily blocked from Resy after the bot queried its API hundreds of times per hour. A tech investor later removed the app, citing the ease with which Instinct could be targeted by phishing attacks. Technology journalist Jasmine Sun warned that “most people’s problems are not software-shaped, and most won’t notice even when they are,” underscoring the limited applicability of such tools.
The growing divide between those who rely on AI assistants for routine tasks and those who have never tried one mirrors the broader industry surprise over data-center backlash. When agents promise to automate administrative drudgery, the environmental and infrastructural costs of scaling compute resources become a point of contention, especially for users who view the assistant as a simple search-engine replacement.
Instinct declined to comment on the testing outcomes or the security issues raised. Despite the risks, the author finds the service valuable enough to keep using it, noting that the agent once warned against opening a phishing-laden email invitation. For busy professionals juggling work and family, the trade-off between convenience and potential data exposure remains a personal calculation.