Instinct Founder Noah Shinn Claims Personal Goal is to Provide This AI Personal Assistant Free for Life to Everyone
Instinct founder Noah Shinn stated on the Invest Like the Best program that his personal goal is to provide this personal assistant free for life to everyone; the platform is approaching an annual transaction volume of about $1 billion, with approximately 50% coming from travel, where users do not pay, and benefiting merchants cover the distribution costs.
The product is invitation-only, with no standalone app; users issue tasks via SMS, voice, or email, and the assistant operates computers and phones in the cloud to complete bookings, hotel reservations, ride-hailing, and calendar entries. Shinn mentioned that marketing expenses are zero, with about 10% daily user growth, and invitations are transferable among users, with transactions on eBay reaching about $300. Within about three weeks, 40% of users linked their personal credit cards, and retention increased to about 80% after linking. More than half of the traffic does not go through iMessage. The company is based in San Francisco under Spear Street Technology, and Shinn previously conducted research at Sierra and participated in the Reflexion paper.
The business model is compared to Apple Pay and American Express: a unified transaction commission, exchanging distribution for merchant access. He cited existing ranges: Shopify around 2.5% to 3%, Amazon up to about 10%, and Apple’s in-app purchases at 30%; airline ticket commissions are usually single digits, while boutique hotels can offer around 30% for customer referrals. He does not promise to achieve 30%, but believes that the travel proportion has proven high commission verticals exist, and he wants to apply the same logic to more industries. The company is currently still in a phase of not charging users and has not monetized transaction volume.
The financing pace has been compressed to a few weeks. In August, a $250 million Series B round was raised at a valuation of about $2.5 billion, led jointly by Index and Benchmark; in September, a $1 billion Series C round was reported, with a valuation of about $10 billion, involving Sequoia, Benchmark, Coatue, etc., bringing total disclosed financing to about $1.35 billion. The demand for computing power is said to double weekly, with Shinn spending about 40% of his time on computing power, with temporary purchases costing three to four times more, and background tasks achieving three, five, or eight times the efficiency on the same computing power. The company’s A/B testing claims it can approach Opus 5 performance at a lower cost.
Use cases include a voice command to book a flight to New York for the same night, seat selection, meal choice, card selection, hotel, round-trip ride-hailing, and calendar entry; evaluations included long-distance check-in and boarding pass push notifications across three airlines. Early users also used it for food ordering, subscription cancellations, concert tickets, wedding planning, and investor data rooms. The target audience is founders, investors, and executives who might hire personal assistants. Invitations start from about 200 friends and family.
In market mechanics, this is about harvesting the software interface into a chat entry point, then taxing transactions after the entry. The buyers are high-net-worth invite users willing to hand over their credit cards to the assistant; Instinct sells order distribution, not subscriptions. Funds flow from hotels and airlines' customer acquisition budgets to the assistant platform. Beneficiaries are boutique hotels and high-commission merchants that can be ordered with one click; the pressured parties are online travel agencies, booking buttons, and apps that rely on advertising for retention. If the commission is not unified, the assistant will encounter the same conflicts of interest as OTAs: those offering higher commissions will be recommended more easily.
Supplementary data: According to the host, if daily growth maintains at 5% to 8% and survives the new computing power rollout period, users could reach about 100 million, which is an extrapolation, not a guideline. The total fee rate on payment tracks is about 2% to 2.5%, which he excludes from pursuit targets. Some analyses estimate that with 80% of hotels in the travel industry and a 10% commission, the theoretical commission on $1 billion in transaction volume would be tens of millions of dollars, provided that users place all orders through the platform rather than redirecting to official websites.
Source: Public Information
ABAB AI Insight
Noah Shinn turns free into a trust strategy: when the assistant knows user preferences better than the users themselves, it cannot rely on ads to sell users. Offering it free for life shifts revenue from users' wallets to merchants' customer acquisition budgets, allowing the product to require credit cards. The three-week 40% card linking and 80% retention after linking are validation metrics for this strategy. No app and entering existing chat windows are meant to eliminate download funnels and transfer distribution rights from app stores to contact lists. The valuation jumped from $2.5 billion to $10 billion, buying the entry tax rights, not realized revenue.
The capital path is to first use computing power to complete orders, then make those orders something merchants must buy as a shelf. Travel is prioritized because hotels already pay high commissions to OTAs, and the assistant only needs to turn "help me book a night" into an order with paid commissions. Airline ticket commissions are thin, while hotel commissions are thick, so the latter segment is what truly sustains the company within the 50% travel volume. A unified commission is necessary to avoid the assistant becoming a talking Expedia. Money flows from Benchmark, Sequoia, and Coatue into GPU queues, then out from boutique hotel marketing budgets. The doubling of computing power weekly necessitates that financing must stay ahead of growth; otherwise, the invitation system will choke the service.
Comparative objects are Apple Pay charging merchants while users are exempt from passwords, and Booking charging high commissions to independent hotels. The difference is that payment only handles funds, while Instinct also handles choices. The industry phase is transitioning from chatbots to order placement: whoever first obtains credit cards and calendars temporarily owns the shelf. OpenClaw, Grok Bot, and Tasklet have been compared by early users, indicating that the window is calculated monthly. If OTAs and super apps make their assistants the default entry point, commission negotiations will immediately decline.
Structural judgment indicates a transfer of pricing power. The mechanism is: once the software interface collapses into a message, search ads and in-app shelves lose exposure, and transactions occur at the moment the assistant confirms on behalf of the user. The commission tax rate depends on how irreplaceable the assistant is to the user, not website rankings. Travel proves that high commission verticals can be absorbed by agents; whether it can expand to the entire industry depends on whether a unified fee rate can prevent merchants from raising prices to buy recommendations. Free is for users, while tax is for the shelf.
ABAB News · Cognitive Law
Free for users is to tax merchants
Only after linking cards does the assistant become the entry point
When the interface collapses into a message, the shelf tax changes hands