Instinct closed a $250 million Series B in August 2026 at a $2.5 billion post-money valuation. Index Ventures and Benchmark co-led the round.

Before the Series B, the company had raised approximately $100 million from Conviction, Greenoaks, Kleiner Perkins, and other investors.

Total funding stands at approximately $350 million. Instinct operates as Spear Street Technology, Inc., headquartered in San Francisco.

Instinct is an autonomous personal AI assistant that connects to a user's email, messaging, calendar, documents, screen activity, audio, and location, then operates their phone and computer to complete tasks. Instead of generating instructions for the user, Instinct navigates websites, fills out forms, sends emails, makes reservations, and reports back.

Users call or text Instinct rather than interact through a separate app or dashboard. For example, a user might text, find a dinner near my hotel tomorrow that works with my calendar and book it for four. Instinct identifies the hotel location, checks the calendar for open windows, searches restaurants, applies saved preferences, books the table, adds it to the calendar, and texts a confirmation.

The assistant connects to Google Workspace (Gmail, Calendar, Drive, Docs, Sheets, Slides, Tasks) through official APIs. With permission, it can also observe screen contents, keyboard and cursor activity, and ambient audio. For services without structured API access, Instinct reads screens, clicks buttons, enters information into forms, and navigates multi-step workflows. Users can provide account credentials and payment information for the agent to sign into services or make purchases.

Users can store sensitive materials in a feature called the Vault, whose contents are excluded from AI model training and used only to serve that individual user. Instinct can also monitor unfinished obligations and contact the user about a dropped email thread, an upcoming calendar conflict, or a stalled planning task.

Early users have used Instinct to book restaurant reservations, airport rides, travel, groceries, and concert tickets; rebook disrupted trips; cancel subscriptions; organize investor data rooms; maintain CRM information; and plan weddings. Its current target users are time-constrained founders, investors, executives, and affluent consumers who might otherwise rely on a human assistant for this coordination.

Instinct has not announced a commercial pricing model. Its current go-to-market is B2C, invitation-led, and venture-subsidized, with the terms of service restricting the product to personal use.

A hybrid subscription and usage model is the most plausible monetization structure. A recurring base fee could cover a monthly allocation of tasks, integrations, and compute, with overage charges or higher tiers for heavy users. Pure flat-rate pricing carries cost risk because each request can trigger dozens of internal model calls, browser sessions, retries, voice transcription, and storage operations, making heavy users more expensive than light users.

Transaction and referral revenue could provide a secondary revenue stream. Instinct sits upstream of commercial intent across travel, dining, transportation, local services, groceries, and ticketing. Merchant commissions or affiliate fees could subsidize consumer pricing, but steering users toward higher-commission merchants would conflict with the assistant's role as the user's representative and could undermine trust.

The cost structure is more variable than conventional consumer SaaS. Cost drivers include foundation-model inference, proprietary agent planning, browser and virtual-device sessions, voice and SMS charges, long-term personal-context storage, integration maintenance, and support for failed or sensitive actions. Gross margins will likely start below mature SaaS norms and improve if the company develops smaller specialized models, caches common workflows, and replaces brittle screen automation with direct API integrations.

The model depends on two related feedback loops. More tasks generate richer context about users' preferences, routines, and relationships, which can improve task completion and encourage users to connect more accounts and delegate more work. Execution data also feeds into model improvement, with successful task traces used to improve reliability across users.

Instinct competes across four converging categories: platform-integrated assistants, independent AI executive assistants, computer-use agents, and ambient-context products. Its breadth differentiates it while exposing it to competitors across each category.

Google Gemini is the most immediate strategic threat. Gemini Spark combines persistent personal intelligence, voice interaction, background execution, proactive briefings, and actions across Gmail, Calendar, Drive, Maps, and more than 40 Android applications. Its connected commerce services include Ticketmaster, OpenTable, Zocdoc, Thumbtack, and Angi. Gemini reported more than one billion monthly active users in August 2026.

Apple Siri AI presents a different challenge. System-level access across iPhone, Mac, Apple Watch, and Vision Pro gives Siri structural advantages in screen awareness, on-device privacy, and cross-app actions through App Intents. Apple does not need to match every Instinct capability. A sufficiently capable, preinstalled Siri could reduce mainstream users' incentive to grant another company access to their communications and screens.

Microsoft Copilot competes less directly in personal errands but limits Instinct's path into corporate workflows through native Windows, Outlook, Teams, and Microsoft 365 integration with existing enterprise identity and compliance infrastructure.

Lindy is the closest independent software competitor, offering autonomous inbox management, meeting follow-ups, calendar coordination, and multi-channel communication through iMessage, SMS, Slack, and email. Lindy cites SOC 2 Type II, HIPAA, and GDPR compliance and states that customer data is not used for model training, giving it a more narrowly defined security model than Instinct's broader data-access model.

Motion bundles AI scheduling, task management, project management, meeting notes, and email into one structured work system. Its deterministic execution and organizational visibility may appeal to users who prefer a product handling 70% of administrative work over an assistant with a broader scope and larger error surface.

Fyxer, Tasklet, and a growing long tail of profession-specific assistants create pricing pressure and fragment demand. Legal, sales, or recruiting assistants can outperform a general agent within their niches by embedding domain-specific rules and integrations.

OpenAI ChatGPT combines voice, personal memory, email and calendar connectors, browser operation, scheduled tasks, and computer use across desktop applications. ChatGPT agent can conduct multi-step web workflows, operate authenticated sites, and recur on a schedule, narrowing the gap with Instinct's persistent-agent model.

Amazon Bee, acquired in 2025, is developing into an ambient personal companion that learns from conversations, email, and calendar data and can turn spoken commitments into actions. Its integration with Alexa, Echo, Prime, Amazon shopping, and home services competes with Instinct's consumer transaction use cases in grocery, household purchasing, and home-service bookings.

Rabbit's Large Action Model operates existing software interfaces through visual UI understanding and user-taught workflows, indicating that computer-use agents can be built without Instinct's broad personal-data access. Open-source systems such as OpenClaw offer self-hosted alternatives for privacy-conscious technical users.

Meta's acquisition of Limitless brings ambient memory into Ray-Ban glasses, WhatsApp, Messenger, and Instagram, creating a platform-level competitor around the interpersonal communication graph Instinct uses for context.

Instinct's initial wedge is a general-purpose personal assistant for time-constrained early adopters, while its product architecture supports expansion into adjacent markets that could increase its addressable base.

Instinct's current terms restrict use to individuals, but coordinated household accounts represent a potential extension. Shared agents could manage family calendars, childcare logistics, household purchasing, travel, home repairs, and eldercare coordination, with granular permissions separating private and shared information.

Many high-friction consumer tasks involve multiple people. Household plans could increase the number of addressable users and the value of each account by extending existing scheduling and communications capabilities with consent management and separate identity boundaries.

Instinct sits upstream of commercial intent across travel, dining, entertainment, local services, grocery, mobility, and healthcare bookings. Google's August 2026 expansion of Gemini connected apps into Ticketmaster, OpenTable, Angi, Thumbtack, Zocdoc, and car rental indicates the breadth of this commerce surface.

Instinct could differentiate by remaining merchant-neutral and optimizing across providers rather than favoring services owned by a vertically integrated platform. Over time, the company could build a developer and merchant platform where service providers expose structured actions, such as rescheduling appointments, requesting refunds, and changing reservations. These APIs could replace brittle screen automation and generate transaction or distribution revenue.

Employers already pay for productivity tools, administrative support, and employee benefits. Instinct could be distributed as an employee benefit for managing travel, scheduling, caregiving, healthcare navigation, and household administration, with the employer paying but receiving no access to personal data.

A separate enterprise product could automate employees' work tasks directly, but would require single sign-on, role-based permissions, audit logs, data residency, and contractual restrictions on model training. The prosumer segment, including founders, investors, recruiters, salespeople, and creators who blend personal and professional workflows, offers a nearer-term path that avoids the full enterprise compliance build.

Instinct is based in San Francisco and denominates services in U.S. dollars. International expansion could add multilingual phone and messaging interactions, local merchant integrations, regional calendars and address formats, and support for country-specific transportation, travel, healthcare, and payment systems. English-speaking markets have the lowest localization costs, while markets with fragmented service providers or less capable incumbent assistants may offer greater differentiation.

Platform displacement: Apple and Google are embedding personal context, screen awareness, connected applications, and cross-app actions into operating systems that already possess user identity, distribution, and extensive personal data, with Gemini alone reporting more than one billion monthly active users in August 2026, which could cause Instinct to lose device permissions or become a thin orchestration layer before it establishes an independent consumer relationship.

Trust concentration: Instinct requires access to communications, audio, precise location, credentials, payment information, and screen activity to deliver its differentiated value, while a single visible unauthorized email, unintended purchase, or prompt-injection attack exploiting content from a malicious email or website could cause disproportionate reputational damage and prompt users to revoke the broad permissions on which the product depends.

Unit-economic uncertainty: Each user request can trigger dozens of foundation-model inference calls, browser sessions, retries, voice transcription steps, and long-running monitoring workflows, resulting in per-user costs far above conventional consumer SaaS despite consumers' preference for low-cost subscriptions, and the company must close this gap before consuming $350 million in venture funding.