Key Takeaways

  • Instinct scaled to $1 billion in annualized transaction volume in roughly six months, beating Stripe, which took 18 months after two years in beta.
  • Founder Noah Shinkeill wants to keep Instinct free for users forever, relying on merchant take rates rather than subscription fees.
  • Platform take rates set the benchmark for agent monetization: Shopify charges 2.5% to 30%, Amazon takes 15%, and Apple takes 30%.
  • Outbound consumer agents will soon face inbound merchant agents, creating automated server-to-server negotiations over pricing and inventory.

The Disagreement

When consumer agents execute purchases, book reservations, and allocate capital, they eliminate the friction that commerce platforms rely on. John Coogan and Jordi Hays look at that friction from opposite sides.

Coogan expects full automation on both ends of the transaction. If buyers deploy agents to snipe restaurant tables and flights, sellers will deploy counter-agents to defend their inventory. As Coogan put it: “I think that the reservation system will be an agent as well. And there might be a country of geniuses in a data center negotiating against another country of geniuses in data center.”

Hays argues that human businesses intentionally build friction to protect value. “I think some of the friction in hospitality is designed to reward people that plan really far in advance and people that develop a real relationship with the business and the people that work there,” Hays noted. If automated agents strip away that relationship, businesses will block bots or introduce dynamic surcharges to restore the barrier.

Who's Right (and When They're Wrong)

Both dynamics will play out, but in different sectors.

Coogan is right in commoditized transactions. Booking flights, ordering consumer packaged goods, and buying hardware do not rely on personal warmth. When price and speed rule, automated negotiation wins. Instinct proved this demand by hitting $1 billion in annualized volume in six months. Coogan pointed out the speed: “For reference, it took Stripe around 18 months to hit 1 billion of annualized transaction volume. Although Stripe also spent two years building in beta before launch, Instinct got to 1 billion transaction volume in closer to six months.”

Yet Hays is right whenever supply is scarce and status-driven. Top restaurants, boutique hotels, and luxury brands do not want frictionless, price-optimized checkout. If every user sends an autonomous bot to book tables at 9:00 AM, the restaurant loses the ability to prioritize regulars. Sellers will retaliate with strict anti-bot verification and deposit walls.

The larger test for personal agents is their unit economics. Shinkeill wants Instinct to remain free for consumers, funded by taking a cut from sellers. “Instinct plans to take a cut of what users buy. Take rates. Shopify is 2 and a half to 30%. Amazon is 15%. Apple is 30%,” Coogan noted. But high per-user inference costs mean a zero-fee consumer agent must capture massive volume to cover server bills.

There is also systemic risk when agents optimize capital. As Coogan noted, the chief economist at Apollo warned that widespread adoption of financial AI agents could trigger bank runs as algorithms automatically move deposits to chase yield.

What to Do With This

If you are building an AI agent, audit your revenue model against your inference costs this week. Calculate your cost per completed action. If your take rate on merchant transactions cannot clear your token spend by at least 3x, introduce paid priority tiers immediately instead of subsidizing user queries.