Instinct raises $1 billion to turn AI assistance into an everyday habit
Last weekend, MTN explored the pleasures of overengineering your life: turning a thirty-second chore into a Saturday-long technology project. Instinct is pursuing the opposite proposition. Its personal AI assistant promises to take everyday tasks off your hands.
On 28 September 2026, the San Francisco company announced a $1 billion Series C round involving Sequoia Capital, Benchmark and Coatue, valuing the business at $10 billion. That followed a $250 million round at a $2.5 billion valuation announced in August.
The rapid increase in valuation reflects investor enthusiasm for AI agents: software designed to carry out tasks on a user’s behalf. Instinct's opportunity is to turn that enthusiasm into a habit, giving people enough useful assistance to keep them returning after the novelty fades.
How Instinct’s personal AI assistant works
Instinct was founded in 2025 by Noah Shinn, who previously worked at Sierra, the AI customer-service company. It operates through Spear Street Technology.
Users can text or call its assistant. Instinct says it connects with applications and devices, including email, messaging and location information, and uses its own phone and computer to act on their behalf. Its website describes arranging airport transport, finding a handyman and following up on unfinished conversations. [4]
The ambition is to reduce the effort of moving between services. Much of everyday administration involves finding information, comparing options, making arrangements and checking that something actually happened. Personal context could help an assistant manage those steps without requiring a fresh explanation every time.
From requests to everyday tasks
Consider an illustrative task: arranging dinner with friends. Finding restaurants is the beginning. Someone must coordinate availability, check dietary requirements, choose a time, reserve a table and circulate the details.
Instinct's September announcement includes a network through which users' assistants can communicate to arrange plans, plus a concierge offering for more involved cases. An assistant that handles the dinner arrangements could remove a string of messages and searches. But “book somewhere nearby” leaves plenty unresolved. How expensive? Accessible to whom? Is everyone happy to pay a deposit?
The value lies in resolving those details with minimal effort from the user. Repeated clarification can erode the time saving; making assumptions can produce an unsuitable booking. Good assistance requires knowing which decisions can be made within agreed preferences and which need approval.
Convenience has a business model
Completing transactions could also support the business. In his September appearance on Invest Like the Best, Shinn discussed earning a share of transactions rather than building an advertising business. Business Insider reported his intention to keep the assistant free for users while merchants pay when it brings them customers and helps complete purchases.
Under that proposed model, a dinner recommendation raises a commercial question: would it reflect the group’s preferences or a merchant’s payment to the assistant? Users would need to understand the relationships shaping their choices.
Business Insider reported on 5 October that some users complained about unsolicited shopping suggestions. The report did not establish how widespread the issue was or whether Instinct earned money from those suggestions. It nevertheless illustrates how proactive assistance can become unwelcome when users have not asked to shop.
A commission model would also need to cover the cost of helping with tasks that generate no purchase. Coordinating friends’ availability, for example, consumes computing resources without necessarily producing income.
How Instinct compares with OpenClaw and Meta Muse
OpenClaw provides a useful comparison. The open-source assistant can run on infrastructure controlled by its user and connect messaging services to AI agents. Its documentation describes tools that can work with files, run commands and send messages, depending on the permissions configured.
That gives technically confident users scope to customise their assistant. It also puts configuration and operation in their hands: connecting services, selecting AI access and managing permissions. Running the software on your own infrastructure does not necessarily keep all data local; that depends on the models and services connected.
Instinct's proposition is to provide assistance as a managed service. Its competitive value therefore includes how much setup and maintenance it removes. Similar capabilities elsewhere do not make convenience worthless, but they raise the standard for what a commercial provider must deliver.
Meta presents a different challenge. Its Muse personal agent, introduced in September, works across applications from a dedicated virtual computer and can be reached through its app or WhatsApp. Meta brings established distribution to the competition.
Privacy and control when AI acts for you
Returning to dinner, a useful assistant should confirm that the restaurant accepted the reservation, explain any cancellation charge and seek approval for spending outside agreed limits. Users need a clear account of actions taken and a way to correct them.
Instinct describes safeguards including separated operating environments and credentials that expire after a short period. These address access and isolation, while accurate interpretation and recovery from mistakes remain separate requirements.
Its privacy policy says access depends on permissions but can include private communications, credentials and sensitive information. It allows some information to support model training. Users can opt out of future training, although information flagged for safety review may still be used. Data received directly through Google Workspace APIs is excluded from model training.
The more context an assistant receives, the more useful it could become. Maintaining trust requires understandable controls over that access.
Can Instinct turn viral interest into lasting demand?
Instinct remained invite-only when Shinn appeared on Invest Like the Best on 28 September. The podcast presented a rapidly growing product, but that account is not an independently audited measure of adoption or evidence of how many people will remain active months later.
Instinct nevertheless addresses recurring work, giving it a plausible route beyond novelty. Its durability depends on making delegation easier, preserving trust and covering the cost of assistance. For the person arranging dinner, the outcome is simpler: a suitable table booked, friends informed and less time spent organising it. Repeating that kind of result is what could turn viral attention into an everyday habit.
Liked this article? You can support our independent journalism via our page on Buy Me a Coffee. It helps keep MoveTheNeedle.news focused on depth, not clicks.