Field report NE-8 / Q6 · sheet /strategy

Corporate StrategyFiled from the territory

Agent Name Service is now a Linux Foundation project for open AI trust

GoDaddy's Jared Sine says AI agents need a DNS-like identity layer. The Linux Foundation's Agent Name Service could verify operators, software and permissions before transactions.

Grid
NE-8 / Q6
Surveyed
November 19, 2026
Reading
3 min read
Correspondent
Priya Raman
GoDaddy chief strategy and legal officer: AI agents still have a trust problem. The solution is waiting
Fig. NE-8 / Q6GoDaddy chief strategy and legal officer: AI agents still have a trust problem. The solution is waiting — viarami / pixabay

Field notes

  1. Agent Name Service was a Linux Foundation project by November 19, 2026.
  2. GoDaddy says it contributed to ANS, a DNS-inspired open standard for AI-agent identity.
  3. ANS verifies four elements: operator, registered software version, ownership or software changes, and permissions.
  4. DNS replaced raw numerical server addresses with readable endings including .com, .gov and .edu.
  5. The proposal warns that proprietary protocols could confine AI commerce to closed commercial ecosystems.

By Priya Raman

By November 19, 2026, Agent Name Service (ANS) was a Linux Foundation project, with GoDaddy among contributors to the open standard designed to give autonomous AI agents verifiable identities before they handle money, data or transactions.

The strategic case is straightforward: AI assistants are moving toward delegated work such as comparing flight prices and managing routine household purchases. Before an agent can act for a person or company, it must distinguish a legitimate service from a convincing imitation.

“The internet needs an identity layer for agents,” GoDaddy chief strategy and legal officer Jared Sine wrote in the commentary.

The stakes reach the structure of the next internet. Sine argues that shared identity could help agents transform the digital economy, while a patchwork of proprietary systems could fragment it into competing platforms and closed ecosystems.

Why does DNS offer a blueprint?

The precedent is DNS (Domain Name System). In the internet’s early days, users had to locate businesses through raw numerical server addresses, much as someone might search a city with no street names or storefront signs. DNS replaced those strings with stable, human-readable names.

Endings such as .com, .gov and .edu organized the web’s destinations. Names including Wikipedia.org and Amazon.com told users where they were going while giving businesses and institutions a unique, durable and universally accessible online presence, often the foundation of a global brand.

DNS also created a trust mechanism. A label such as Amazon.com could function as a trusted commercial identity because users could determine where it led; without that system, they had little basis for separating Amazon’s servers from a convincing imitation run by fraudsters.

What can ANS actually verify?

ANS would give agents a public way to identify one another, verify who operates them and confirm what they may do. Sine says the model can follow DNS because the existing naming system already performs a comparable function for websites and email services.

The standard is designed to let every business verify four elements:

  • The business operating an agent.
  • Whether the agent’s software matches its registered version.
  • Whether ownership or the software itself has changed.
  • What the agent is authorized to do.

The practical target is pre-transaction verification. An AI travel agent should confirm that it has reached an airline’s legitimate booking system, not a polished copy. A shopping agent should verify the merchant, payment processor and delivery provider before placing an order.

Can a legitimate agent still be manipulated?

ANS addresses a subtler risk than direct impersonation. A supplier could buy a company’s AI procurement agent, retain the same software and domain name, then nudge the system to recommend the supplier’s products. The buyer might see no obvious change.

Yet the agent would no longer act entirely in the customer’s interest. Sine says businesses without a shared identity and verification layer may have no reliable way to detect that shift or confirm that agents perform the work their owners expect.

Do private agent networks solve the trust gap?

Some companies are already developing communication among agents, particularly in payments and logistics. Sine argues that interoperability within a handful of commercial ecosystems is not equivalent to a global open standard that machines can follow across networks.

Left to proprietary protocols, machine identity would confine the agent economy to private silos. One company’s agents would work best with its preferred merchants, payment providers and service platforms, limiting the reach and choice available to customers elsewhere.

The internet’s current architecture grew from shared infrastructure that no single company controlled. Sine’s proposed standard applies that principle to AI agents: common identity rules should sit above individual platforms, much as DNS sits above websites and email services.

“Without a common identity layer for AI agents, the promise of the agent economy will give way to confusion and fraud,” Sine warned. “And we’ll still be spending hours trying to find the best deal on plane tickets.”

via fortune.com/feed/ (Source)

Share this article:

More from Priya Raman

Priya Raman

Show full bio

Senior reporter covering technology and product news in corporate strategy and chief-executive leadership at CEO Frontier.

460 articles

Filed under

Adjacent sheets

« Previous articleNext article »

End of sheetNE-8 / Q6 · November 19, 2026