Instinct AI: A Text-Based Personal Assistant That Challenges Meta Muse and ChatGPT Dots

Instinct AI is taking a different approach to personal AI assistants by handling everyday tasks through text messages, WhatsApp, and email instead of requiring users to manage another app. The startup wants its assistant to do more than answer questions, offering help with appointments, travel bookings, purchases, email correspondence and other administrative work.

Founded by former Sierra researcher Noah Shinn, Instinct has attracted significant investor interest. The company reportedly raised $1 billion in September 2026 at a $10 billion valuation, even though its assistant remains an invitation-only service. It is entering an increasingly competitive market alongside Meta Muse and OpenAI’s ChatGPT Dots.

How Instinct AI Works

Instinct’s main advantage is its familiar interface. Users describe what they need in plain language, and the assistant tries to complete the necessary steps using connected accounts and online services. It can research options, contact businesses, manage correspondence, and request clarification when a task needs more information.

According to the official Instinct website, the assistant can help order groceries, arrange gifts, dispute bills and plan trips. Its design removes much of the complexity of conventional automation tools, letting users delegate tasks without configuring individual workflows.

The company has also introduced dedicated email addresses for its agents. These let Instinct communicate with businesses, manage confirmations, and handle certain account-related activities without requiring every interaction to go through the user’s personal inbox.

Instinct Can Make Calls and Join Group Chats

Instinct has expanded beyond messaging with a phone-calling feature called Instinct Concierge. The capability lets the assistant contact businesses about reservations, appointments, and other requests that cannot be completed entirely online. The feature initially launched through an early-access program.

The company has also introduced group-chat assistance to coordinate trips, events, and shared plans. Participants can communicate with the same AI agent, including people without individual Instinct accounts. The system keeps personal assistant information separate from group conversations unless permission is granted.

These additions make Instinct more useful for tasks involving other people, although reliable execution remains essential. A phone conversation or group booking can introduce complications that an automated system can’t easily anticipate.

Early Testing Shows Promise but Also Mistakes

Independent testing suggests Instinct can deliver practical results, but it is not consistently reliable. In a September evaluation, WIRED journalist Zoë Schiffer used the assistant to handle financial disputes, restaurant reservations and email-related tasks.

Her published experience included recovering $550 and identifying a suspicious email. However, the assistant also caused an unnecessary $64 expense, showing how an agent can do useful work while still making financially consequential mistakes.

This is Instinct’s biggest limitation. Unlike a chatbot that provides information, an autonomous agent may interact with businesses and take actions involving real accounts or payments. Users should therefore review important transactions, confirm bookings and limit access to sensitive information where possible.

Instinct vs Meta Muse and ChatGPT Dots

Instinct competes with Meta Muse and ChatGPT Dots, but its strongest distinction is simplicity. Instead of making a separate app the centre of the experience, Instinct lets users request assistance through familiar messaging channels.

Muse benefits from Meta’s broader consumer ecosystem, while ChatGPT Dots emphasises connected tools and more extensive personal and professional workflows. Instinct focuses more narrowly on completing everyday errands with minimal interaction.

None of these approaches guarantees better task completion. Their usefulness depends on the services they can access, how reliably they perform actions and how much control users retain over important decisions. There isn’t enough standardised, independent testing to identify a clear overall winner.

Privacy, Pricing and Availability

Instinct needs access to personal information and connected accounts to perform many tasks. That raises legitimate questions about permissions, data handling, and the potential consequences of mistakes. The company faced criticism over its early privacy terms and subsequently revised its documentation.

The service remains invitation-only and is currently free for accepted beta users. Instinct has not confirmed its long-term subscription pricing or full monetisation strategy. The company has also explored personalised recommendations through Instinct Selections, which could create additional commercial opportunities.

For prospective users, the lack of permanent pricing and broader availability makes it difficult to assess the service’s long-term value. Its capabilities are promising, but the costs, reliability and privacy implications will become clearer as access expands.

Final Verdict

Instinct AI offers a practical vision of personal AI assistance. Its messaging-based interface, email capabilities, phone calls and group coordination tools could reduce the time people spend on repetitive administrative tasks. The approach is especially appealing to users who want assistance without learning another complicated application.

However, early testing has exposed mistakes, and the service still faces questions about privacy, reliability and future pricing. Instinct is worth considering for early adopters comfortable supervising automated tasks, but it is not yet a dependable replacement for a human assistant. Its success will depend on whether it can complete everyday work accurately and consistently as more people begin using it.