Key Takeaways

  • Instinct connects directly to personal email and SMS to execute autonomous web tasks for individual users.
  • Relentless automated booking attempts triggered immediate user account bans on platforms like Resy.
  • In one test case, the software contacted hundreds of people across broadcast networks to retrieve seconds of US Open fan cam footage.
  • Outbound agent volume will inevitably force receiving platforms and individuals to install inbound AI gatekeepers.

The Agent That Refuses to Stop

Consumer software used to wait for your next click. New autonomous assistants do not wait around. Instead, they take a broad goal, link into personal text messages and inboxes, and execute until they get a result or hit a brick wall.

On a recent TBPN episode, John Coogan and Jordi Hays examined the rise of Instinct, an autonomous assistant built directly for consumer workflows. Hays highlighted how the tool created an incredible user experience while simultaneously breaking platform rules. In one instance, a user wanted a difficult dinner table. Hays noted, “there was one report of an individual who got banned from resi because they were absolutely spamming to get reservations to restaurants.”

In another scenario, a tennis spectator spotted themselves on screen at the US Open and asked the software to find the broadcast clip. To the user, the interaction felt like magic: type one sentence, get the video. But behind the scenes, the bot ran wild. As Coogan described it, “what Instinct was doing in the background was being absolutely relentless and it seems like it was reaching out to hundreds of people aggressively, you know, following up.”

For the consumer, a personal agent that pesters hundreds of television staff members to track down five seconds of tape is a massive superpower. For the people on the receiving end, it is indistinguishable from a distributed denial-of-service attack on their attention.

The Inbound Filter Arms Race

When software reduces the marginal cost of outbound communication to zero, outbound volume explodes. Web platforms and service workers cannot manually review hundreds of automated follow-ups per transaction. That imbalance guarantees that existing anti-spam defenses will fail.

Coogan argued that the only viable solution is symmetric automation. Humans cannot filter agent traffic, so agents must talk to agents. As Coogan explained, “you could imagine if the other side has agent intermediation and filtering and can process things that are spamlike, but there's actually a reasonable action to take that could make it through a filter and not actually annoy anyone on the other side.”

Right now, platforms react with blunt instruments: IP bans, account deletions, and hard captchas. Resy bans the user because a brute-force reservation loop degrades the platform for everyone else. But as consumer agents spread to travel, customer support, and local services, hard blocks will alienate real customers. The next generation of consumer infrastructure will not just build agents to do your chores. It will build gatekeepers to absorb everyone else's bots.

What to Do With This

Audit your product's public contact forms, reservation endpoints, and inbound support queues this week. Identify where an automated agent could submit 200 variations of an inquiry without hitting a rate limit or a verification gate. Add rate limits based on authenticated identity rather than simple IP addresses before autonomous consumer agents flood your queues.