Back in February 2026—an eternity in artificial intelligence (AI) cycles—I wrote that the sudden rise of OpenClaw was the "Napster moment" for agentic commerce: the raw, unregulated early signal that personal software agents would soon live embedded inside the messaging apps we already use all day.
Eight months later, we seem to have moved into the "Spotify moment," where ambient, personal AI agents are becoming a safe and frictionless reality for the mass market:
- Grokbot, SpaceX’s personal agent, launched in August and quickly earned early tech-sphere praise
- Meta rolled out Muse, their WhatsApp-embedded AI agent, eclipsing 5 million downloads in a mere 22 days
- OpenAI announced Dots on Sept. 29, an ambient assistant embedded within everyday messaging and inboxes.
- The most unexpected trend is the explosive rise of Instinct—a nimble, ambient agent operating straight out of WhatsApp and iMessage that has become the breakout surprise of the agentic era.
A new kid on the block
In a recent interview, Instinct founder Noah Shinn disclosed some mind-blowing numbers: Instinct has already crossed $1 billion in annualized transaction volume and a growth rate of 10% day-over-day with zero marketing spend, operating on an invite-only model.
The headline number of $1 billion is impressive, but the real revelation is where the money is actually moving. Shinn shared that 50% of their transaction volume is due to travel alone.
Let that sink in. That is $500 million travel sales flowing through text prompts and voice notes.
Out of curiosity, I asked AI to run the compounding math and the exponential scale felt absurd: If Instinct would be able to sustain its 10% day-over-day growth from its current $500 million GMV in travel, it would take roughly one month to match the $9.8 billion gross booking volumes of MakeMyTrip last year, and barely 62 days to rival the $186 billion of the largest travel company on earth, Booking Holdings.
Even treated purely as an intellectual thought experiment, the takeaway is hard to ignore. Few purchases carry the friction and the emotional weight of booking a trip. It is expensive, logistically fragile and personal.
Yet Instinct's rapid success in travel suggests that when given the choice, a fast-growing cohort of travelers is happily delegating travel planning and potential purchase to a pure chat interface that lives inside their everyday messaging channels.
The trust flywheel
Booking Holdings CEO Glenn Fogel has repeatedly argued that while generative AI and autonomous agents excel at search and recommendation, trust and execution remain the true moat separating an AI interface from an actual travel platform.
Instinct's founder claims to have cracked the trust challenge through a smart behavioral mechanism. Rather than forcing users into an all-or-nothing trade-off on day one, the architecture relies on gradual data sharing and escalating utility that compounds over time:
- Low friction is the starting point: Traditional e-commerce players are notorious for front-loading friction. Download an app, and you are instantly bombarded with permissions pop-ups: location access, contacts, notifications and payment setups. Instinct flips this dynamic by establishing a gradual, user-directed pace with early lightweight interactions limited to simple text queries or basic calendar scheduling.
- Trust is earned in small increments: Persuading a rational human to let an autonomous agent transact on their behalf requires consistent, ambient proof over time. Instinct measures this progression using the behavioral milestones of "time to first credit card”. According to Shinn, once a user observes three weeks of flawless, low-risk execution, 40% of users voluntarily grant the agent direct access to a personal credit card.
- The 80% retention moat: The moment an agent secures the user’s payment details, Instinct's user retention leaps to 80%. Once the agent becomes the trusted custodian of a traveler's wallet, a consumer has no real reason to switch back to the traditional booking funnel as long as the purchasing experience remains flawless.
- Compounding proactivity reinforces trust: The beauty of the flywheel is that data depth directly expands agent autonomy. The more ambient context the assistant absorbs, the less input it requires to deliver high-impact results and strengthen the trust layer.
With users live location, seat configurations preferences, hotel tastes, city locations and payment rails registered, multi-step travel can condense into a single command:
"I need to be in New York tonight."
With that single voice note, the agent can cross-reference flight schedules, confirm preferred cabin seating, book a tailored boutique hotel room and schedule an Uber to meet at the arrival gate.
Show me the money
If the consumer pays nothing and there are no sponsored search results, how does an ambient assistant that costs nothing to use actually build a business? Instinct's founder's answer is clear: Leave the user’s wallet alone and charge the merchants on the other side of the counter. It is a take-rate model designed to preserve pure user alignment.
During the Meta Connect 2026 keynote, Mark Zuckerberg, who built his empire around eyeball monetization, signaled an identical economic roadmap for Meta’s personal agent, Muse. "We expect that over time we will profit by taking a small fee from transactions," he said.
All roads seem to lead to the same destination: AI agents are transforming into high-intent transaction networks. Instead of fighting for ad clicks, they will monetize the checkout itself, charging a distribution fee to suppliers and travel intermediaries in exchange for delivering verified, zero-friction bookings.
The balancing act between human browsers and autonomous agents
The transition to agentic travel will not happen overnight. During the e-commerce revolution in the late 1990s, it took years of security guardrails and gradual adoption before consumers felt comfortable typing their credit card numbers into a travel websites checkout page. Yet once that threshold is crossed, the behavioral lock-in is real.
The real dilemma is not whether ambient assistants will instantly render multibillion-dollar ad channels obsolete or wipe out hard-earned loyalty programs.
For the foreseeable future, travel brands must continue serving human travelers who still find comfort in browsing visual interfaces while simultaneously re-engineering their backend pipes and forging commercial alliances with AI gatekeepers to capture the growing volume driven by autonomous agents.
About the author...
Mario Gavira is CMO of
Travelier, as well as an angel investor.