After Meta's AI agent, Muse, made waves in the world of artificial intelligence just a month ago, Meta has taken another major step. To eliminate the chaos caused by AI bots in the business world, Meta has joined forces with major companies.
Meta, Walmart, Stripe, and several other companies have created a new open standard called the "Personal Agent Protocol." Its purpose is to set rules for how AI agents interact with businesses.
Meta launched its personal agent, Muse, in early September 2026. It went viral immediately, surpassing ChatGPT to reach number 1 on the Apple App Store. It remains at the top today. People are using it for shopping, booking appointments, and planning travel.
But this is where a new problem arose. When thousands of AI bots began visiting companies' websites, purchasing goods, and collecting data, companies were at a loss to distinguish between real humans and bots.
In response, Amazon blocked Meta's agents and Perplexity's AI agents from its website. Amazon alleged that these agents were scraping websites. Amazon even sued Perplexity in November, alleging that the company was secretly using its AI agents to collect data from Amazon's website without permission. Perplexity called this a "bully tactic."
Walmart took a different approach. It allowed AI agents but installed human verification buttons that prevented even real humans from logging in. Users complained on X that Walmart had blocked humans in an attempt to stop bots.
To end this chaos, companies like Meta, Walmart, Stripe, Sierra, Genesys, Rocket, NiCE, and Decagon have come together. The idea for this protocol was conceived by Bret Taylor. Who is Bret Taylor? He is the former co-CEO of Salesforce, chairman of OpenAI, and co-founder of enterprise AI startup Sierra.
In an interview with CNBC, Bret Taylor said, "Companies need to know whether this is a personal agent or a real human. For many companies, the risk is that a random bot, not acting on behalf of a human, will get this service. Until such a standard is established, it's complete chaos."
This protocol will work similarly to how we see "Sign in with Google" or "Login with Facebook" today. Bret Taylor helped create this system when he was Facebook's technology chief.
1. Identification: Businesses will know this isn't a random spam bot, but a trusted agent acting on behalf of a real user.
2. Security: David Singleton, VP of Engineering at Meta Superintelligence Labs, stated that agents require credit card numbers and personal details for shopping. This protocol will give customers complete control and visibility over this data exchange.
3. Transparency: Companies will be able to see what the agent is doing on their site.
Today, there are three major obstacles to doing business with AI agents: scraping, fraud, and CAPTCHA. Companies fear that bots will copy all of their website data and place fake orders. Therefore, they implement CAPTCHA and human checks, which hinder the AI agents' work.
If the Personal Agent Protocol is successful, these barriers will be removed. A user will say to their Muse-like agent, "Order me milk from Walmart," and the agent will securely authenticate and place the order, without scraping the website.
Meta is the biggest name in this initiative. But it's worth noting that OpenAI and Anthropic have not yet joined this group, even though Bret Taylor himself is the chairman of OpenAI. This suggests that there's still a tussle among major companies regarding the future standards for AI agents.
Experts believe that in the next two years, 50% of internet traffic will be generated not by humans, but by their AI agents. If regulations aren't established now, each website will have to build separate barriers to combat bots. Meta's protocol could serve the internet as UPI did for payments—a common rail.
If successful, this protocol could transform the entire approach to online shopping, customer support, and business automation.
No comments:
Post a Comment