Open Intelligent Agents in Business: The End of the Advertising Era

Welcome to the new era of open intelligent agent commerce.

Written by: Sam Ragsdale

Translated by: Chopper, Foresight News

The era of intelligent agent commerce has arrived. The ACP and UCP protocols promise one-click checkout within ChatGPT and Gemini. Soon, hundreds of millions of consumers worldwide will easily find better products, merchants’ conversion rates will significantly increase, and platforms can earn 5%-10% commissions.

However, built-in checkout in ChatGPT is just an incremental improvement. It cannot reshape society like the internet did in the early 21st century. Only open intelligent agent commerce can achieve that.

Why Wall Gardens Were Ultimately Overturned by Open Protocols

To understand why, we must go back to the 1990s.

At that time, there were two completely different, competing “forms” of the internet.

America Online Model: unified monthly subscription payments, integrated email, weather, and other curated content, eventually encompassing the entire Time Warner content library.

Open Protocol Model: HTTP, DNS, HTML, and a browser called Mosaic.

Compared to the former, Mosaic seemed absurd at the time: websites were few, no need for search, alphabetical indexing was enough. Eight years later, AOL merged with Time Warner for $350 billion, and the market believed curated content was the future.

But soon after, Mosaic and open protocols ultimately prevailed, and human civilization officially entered the digital age.

Why? Let’s hypothesize what would have happened if the closed garden had won back then.

In 2004, Mark Zuckerberg wanted to create Facebook, but first had to negotiate distribution deals with AOL. Two Stanford students wanted to build a web index and needed permission from CompuServe. Someone wanted to sell books online from their garage and had to propose to Microsoft’s MSN content team.

They would have been turned down: “Go back to school, kid.” If that had happened, the digital economy we take for granted today might not exist at all.

The core of open protocols is the absence of gatekeepers. Anyone with a server and a domain name can reach the entire web. Innovation occurs at the edges; the center struggles to keep up, ultimately creating one of the largest waves of wealth in human history. This is the fundamental logic of capitalism: disruption always comes from the margins.

Returning to 1997. Tim Berners-Lee and Marc Andreessen were developing related protocols and browsers. At that time, setting up a server cost hundreds of thousands of dollars, and no one understood why content servers needed to respond to unknown users: high costs and no clear revenue.

They designed a status code called “402,” which the server could return to users with a prompt: “Pay to access this content.” But digital payments were still impossible: PayPal hadn’t been born yet, credit card fees per transaction were often tens of cents, which was too high for small transactions of a cent.

Even so, the internet rose.

Google found an alternative business model: advertising. In traditional media, content creators and users directly establish economic relationships. Google relied on broadcast economics, introducing third-party advertisers who paid to sustain content supply and user reach.

Brilliantly simple. Creators didn’t need to build user relationships; they could monetize just through traffic and attention. Google captured this flow of funds, acting as an intermediary between advertisers and content creators, taking commissions at will.

Small payments were thus sidelined. Open-source software flourished, cloud computing revolutionized, and server hosting costs plummeted by a hundredfold. Google became the staunchest supporter of a free, open internet: the more users searched, the more Google earned. It invested hundreds of billions of dollars to make the web faster, cheaper, and ubiquitous.

AI Intelligent Agents End Traditional Advertising and Wall Barriers

Fast forward to the 2010s, the industry stagnated.

Interest rates continued to fall, capital became more conservative, and the foundational innovations of the internet lost their edge. Major wall gardens tightened their grip on users and accumulated strength.

In 2022, ChatGPT launched, restarting the wave of global transformation. Large language models not only produce results but also integrate multi-source information to generate concise summaries, often without directly fetching native content.

By GPT-4, the trend was clear: intelligent agents would become the next core. They operate computer tasks as skillfully as humans, with lower costs and higher efficiency.

At this moment, the fundamental economic logic of the internet is being rewritten.

From 1997 to 2024, the core of internet profit was attention: humans browsing web pages are easily distracted by ads, and platforms monetize fragmented attention. But large language models and intelligent agents are immune to such distractions.

A highly ironic cycle. Advertising created the free, open internet; massive web data trained large language models; and ultimately, these models will end advertising itself.

After GPT-4’s release, Stack Overflow traffic plummeted 75%, and tech news sites saw a 60% drop in visitors. Tech early adopters are just the beginning; this revolution will sweep through all information scenarios online.

Built-in checkout in ChatGPT is insignificant. The internet is fundamentally a human civilization square, and the old commercial contracts have already failed.

Some corners of the internet still rely on differentiated content to resist crawling—classic wall gardens like Facebook, TikTok, LinkedIn. Thousands of high-paid engineers guard these areas around the clock, blocking bots.

But now, the wall defenses have collapsed. AI intelligent agents can perfectly mimic human online behavior, fooling all defenses. Over the next decade, countless hype schemes claiming to crack this will emerge, venture capital will pour in, but there are no truly effective solutions. Just like fighter jets rendered ground fortresses obsolete, the same will happen here.

The Era of Open Intelligent Agents

What’s next? The answer is open intelligent agent commerce.

Built-in checkout in ChatGPT is like America Online in the era of intelligent agents: curated directories, closed-loop walls, optimized experiences. To sell products through it, merchants would need months of business expansion, rigorous legal documents, five-year plans, substantial revenue, a large user base, and a story that makes headlines like The New York Times, thrilling shareholders.

Open intelligent agent commerce is the current equivalent of the HTTP protocol. A simple, universal standard that allows intelligent agents to pay-as-you-go for everything: data, cloud hosting, communication services, and countless new scenarios yet to be born.

Two pioneers have already implemented solutions: Coinbase launched the x402 protocol, and Tempo partnered with Stripe to introduce the mpp protocol. After 28 years since the birth of the “402” status code, a feasible implementation has finally arrived. Modern blockchain stablecoins with fixed per-transaction fees under 1 cent have perfectly solved the cost problem that killed small payments in 1997.

Intelligent agents that can only purchase from pre-approved whitelist merchants are like employees with limited vendor cards; those connected via open protocols are like entrepreneurs holding free bank accounts.

No need for negotiations, no need for list reviews—only simple, permissionless standards.

These protocols focus on two main points:

• On the agent side: How to make payments easily?
• On the merchant side: How to verify payments?

Large language models excel at calling tools they’ve never seen before. Starting with Claude 4.5+ and Codex 5.2+, intelligent agents can discover APIs, read their patterns, and use them correctly without prior training.

Current discussions focus on skills. They are essentially natural language programmable modules, freely assembled and combined. Non-technical entrepreneurs can automate tasks with everyday language:

• Buy pizza from a nearby highly-rated place and check delivery status every 10 minutes.
• When the driver is five minutes away, turn on the porch light.
• If delivery arrives within 30 minutes, tip the driver $5.

No coding, no programming skills needed. The intelligent agent understands the intent, generates native code in real-time, executes, and then destroys itself. Programming is no longer a required professional skill; natural language mastery suffices.

Skills are indeed effective, but only a transitional form. They are the first intuitive manifestation of the intelligent agent’s ability to call unfamiliar tools. They require dedicated personnel to write, publish, review for safety, and update, and must be pre-installed—inefficient and cumbersome.

The hype around skills masks a deeper revolution: intelligent agents can combine various capabilities in unprecedented ways.

Pizza is just a simple example; real business scenarios are far more transformative. For instance, supply chain management agents for small businesses can detect a 15% price hike from tariffs on packaging suppliers, autonomously find three local alternatives, request samples, negotiate bulk prices, and switch suppliers—all before the owner’s morning jog.

No API integrations, procurement teams, or bidding processes needed—only a balance account and open protocols.

Intelligent agents can make payments and create content, but they currently cannot find what they need on their own.

The remaining challenge is exploration: “How do brokers find what they want to buy?” and “How do merchants introduce their services to brokers?”

The industry has already built a universal search and registration ecosystem. Service providers only need to register nodes at x402scan.com or mppscan.com to connect to the global open protocol intelligent agent network, enabling standardized supply-demand matching and small-value automatic settlements.

In 1997, the internet lacked business models; no one understood why servers responded to strangers. Open protocols and advertising broke this deadlock, ushering civilization into the digital age.

By 2026, advertising as a makeshift solution will fade away. Open protocols and a 28-year-old status code are about to replace it.

Welcome to the new era of open intelligent agent commerce.