Most corporate leaders ignored the opportunity. Some dismissed it as a gimmick. Others never seriously considered it.

This summer, they have another chance. 

ICANN — the organization that coordinates the internet’s addressing system — is accepting applications through August 12 for branded internet endings. Instead of only building their online presence around “.com,” companies can apply to control suffixes that match their own brands, such as “.google” or “.amazon.”

The last application round took place in 2012. Once the current window closes, another opportunity may not arrive until well into the next decade. 

Not every company will see the need for its own “dotBrand.” But it represents a form of digital infrastructure that could become more important in the years ahead, as AI systems, automated interactions, and synthetic content reshape the internet.

For three decades, companies have built their digital identities on infrastructure they do not fully control. They register domain names, defend against impersonators, buy up misspellings, monitor fake websites, and try to steer customers toward legitimate channels. 

The basic bargain has mostly worked. Consumers type in a familiar web address, look for a company name, and assume they have arrived in the right place.

But that bargain is under strain. 

Artificial intelligence is changing how people search, shop, communicate, and transact online. Increasingly, the “customer” a business encounters online may not be a person at all. It may be an AI agent researching a purchase, booking a reservation, comparing suppliers, or handling a customer-service request on someone else’s behalf. Google CEO Sundar Pichai recently predicted that traditional web search will evolve into an “agent manager,” where AI systems run multiple background threads to complete complex transactions for users.

Meanwhile, synthetic content, fraudulent websites, and convincing impersonations are becoming easier to produce at scale.

In that environment, proving who is real becomes more difficult — and more important.

A dotBrand offers one way for companies to create a more controlled and authenticated digital environment. Unlike a conventional web address, which sits inside a shared naming system, a branded top-level domain gives a company authority over an entire namespace. Every address ending in that suffix can be governed by the company itself.

That distinction may sound technical or semantic. But it’s significant.

A bank could reserve its suffix exclusively for customer-facing services, making every address ending in “.bankname” an authenticated destination. 

A retailer could unify e-commerce, loyalty rewards, and customer support under a single branded namespace that customers — and eventually AI systems — know belongs to the company. 

A manufacturer could organize product information, warranties, and partner portals under a single authenticated architecture. 

Over time, AI systems may come to rely on these signals when deciding which sites, services, and commercial interactions are legitimate.

A dotBrand will not solve every problem of trust online. It will not eliminate fraud, stop phishing, or replace cybersecurity. But it could become part of the infrastructure companies use to prove who they are in an internet where that question is becoming more complicated.

The comparison to 2012 is instructive. When ICANN last opened applications, many companies saw little reason to act. The internet revolved around search engines, websites, and human users. Social platforms were rising. Mobile commerce was still maturing. The idea that a company might need its own branded internet suffix seemed remote.

Today, it doesn’t. Companies are investing heavily in AI, cybersecurity, verification, brand protection, and digital identity. They are preparing for a world in which more online interactions happen through automated systems and more customers need assurance that what they see is genuine.

Against that backdrop, a dotBrand is less a marketing flourish than a long-term option on trust and control.

That does not mean every company should apply. Some will conclude the investment is unnecessary. Others will decide the operational burden is too great. Many will have better uses for their capital and attention.

But they will need to actively make a decision. Major changes to internet infrastructure do not come around often. 

Companies routinely spend years debating acquisitions, factory investments, and technology platforms. Opportunities to acquire a permanent piece of internet infrastructure are considerably rarer.

The strategic error is not saying no. It’s discovering years from now that a once-in-a-decade decision came and went without anyone in the boardroom realizing it was theirs to make.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.