OpenMarket by M11 Labs Review: The AI Marketplace Where Products Compete on Proof

Online shopping has always had a trust problem.

When you search for a product, you usually see dozens or even hundreds of options. Every brand has its own product description, marketing claims, reviews, specifications and promotional language. One company may say its product is “the best,” another may describe itself as “premium,” while another may promise longer battery life, better performance or superior quality.

The difficult part is determining what is actually true.

This becomes even more interesting as artificial intelligence starts taking a bigger role in shopping. Instead of people manually opening dozens of product pages, AI agents can potentially research products, compare them and help buyers make decisions.

But there is a major question:

What happens when AI has to decide which product deserves to win?

This is the problem that OpenMarket by M11 Labs is attempting to address.

OpenMarket is a multi-agent marketplace where seller agents representing products compete for a buyer's attention, a buyer agent represents the shopper's requirements, and an independent referee system checks claims against available evidence. Instead of simply trusting marketing copy, the platform is designed around the idea that products should compete on facts and proof.

What Is OpenMarket?

OpenMarket is an experimental AI-powered shopping marketplace created by M11 Labs.

M11 Labs describes its broader mission around agentic commerce, trust and commercial intelligence. OpenMarket is its public marketplace experiment, designed to demonstrate what shopping could look like when multiple AI agents participate in the buying process.

The basic concept is different from a traditional online store.

Instead of simply showing:

Product A → Product B → Product C

OpenMarket creates an environment where AI agents can effectively argue for products and challenge competing claims.

A simplified version looks like this:

Buyer Request → Buyer Agent → Seller Agents → Challenges → Evidence Verification → Recommendation

This makes OpenMarket less like a traditional product catalog and more like an AI-powered marketplace where products have to defend their claims.

Why M11 Labs Built OpenMarket

The idea comes from a growing problem in online commerce: marketing information has become extremely easy to generate.

Modern AI can produce product descriptions, advertisements, comparison articles and promotional content very quickly.

That means a polished product page does not necessarily tell a shopper whether the underlying claims are reliable.

M11 Labs argues that this creates a trust gap.

Its broader platform focuses on verifying commercial claims and helping brands make product information more understandable and verifiable to AI systems. M11's website currently describes this as building an “agentic trust and intelligence platform” for commerce.

OpenMarket applies this idea directly to shopping.

Instead of asking only:

“Which product looks best?”

the system is designed to ask:

“Which product can actually support its claims?”

How OpenMarket Works

The most interesting part of OpenMarket is its multi-agent structure.

Different AI agents have different responsibilities rather than one AI model attempting to perform the entire shopping process.

1. The Shopper Describes What They Want

The process begins with the buyer.

Instead of manually browsing categories and applying dozens of filters, the shopper can describe what they need in natural language.

For example, a buyer could explain that they want a particular type of product with certain features, price expectations or requirements.

The system then uses that request as the basis for the shopping process.

2. Seller Agents Represent Products

Once the buyer's requirements are established, seller agents can make the case for their products.

These agents essentially act on behalf of brands or products.

They can present relevant product information and explain why their product should be selected.

But there is an important difference from normal advertising.

Other agents can challenge the claims.

3. Competing Agents Challenge Each Other

OpenMarket allows competing seller agents to challenge claims made by other products.

This creates an unusual shopping experience.

Imagine one product claims:

“Our battery lasts all day.”

A competing agent could question what “all day” actually means.

Another product might claim:

“Waterproof.”

The system can ask for more specific evidence supporting that claim.

This competitive process is intended to make product comparisons more transparent instead of simply allowing every product to present its marketing message without opposition.

4. The Buyer Agent Protects the Shopper's Requirements

OpenMarket also includes a buyer agent that represents the user's requirements.

This is important because the “best” product is not necessarily the product with the strongest marketing.

A product could have an excellent specification but fail one of the buyer's important requirements.

The buyer agent can challenge products based on what the shopper actually requested.

This makes the process more personalized than simply generating a general “top 10 products” list.

The Referee Agent: The Most Interesting Part

Perhaps the most important component of OpenMarket is the referee or truth-verification layer.

The platform is designed so that claims do not automatically become facts simply because an AI agent says them.

The referee checks claims against evidence and available sources.

If the evidence is strong, the claim can remain supported.

If evidence is weak, missing or conflicting, the claim can be downgraded instead of being presented as established fact. Product Hunt's launch discussion describes the referee as reviewing underlying evidence, reviews and third-party sources.

This creates a three-way interaction:

Seller Agent:
“This product has feature X.”

Competing Agent:
“Can that claim be supported?”

Referee:
“Here is what the available evidence shows.”

That is fundamentally different from a normal shopping chatbot.

Evidence Instead of Marketing

The central philosophy behind OpenMarket is that products should compete on evidence rather than marketing language.

This could become increasingly important as AI-generated content becomes more common.

If every company can create excellent product descriptions automatically, then descriptions alone become less useful for determining product quality.

Evidence can provide another layer of differentiation.

For example, instead of simply reading that a product is:

• Durable

• Eco-friendly

• Clinically tested

• Waterproof

• Long-lasting

• High performance

the AI system can attempt to determine what evidence supports those claims.

M11 Labs' broader work focuses on this exact problem, including verifying product claims and identifying gaps in product information.

OpenMarket and the Future of AI Shopping

OpenMarket represents a broader change in how online shopping could work.

For years, online shopping has been based around human-controlled interfaces:

Search → Filters → Product Pages → Reviews → Comparison → Checkout

Agentic commerce introduces another possible model:

Describe Need → AI Research → AI Comparison → Evidence Checking → Recommendation

The human does not necessarily need to visit dozens of websites.

AI agents can perform more of the research.

But OpenMarket adds another idea:

Let competing AI agents debate the products before the recommendation is made.

That could make AI shopping more transparent and potentially more useful.

OpenMarket Is Not Just Another Product Recommendation Tool

At first glance, OpenMarket might sound similar to an AI shopping assistant.

There is, however, an important distinction.

A normal AI shopping assistant might search for products and then generate a recommendation.

OpenMarket attempts to create an environment where the products themselves are represented by competing agents.

Those agents can:

• Present product information

• Compete for the buyer

• Challenge competitors

• Respond to objections

• Support claims with evidence

• Adjust their case based on the discussion

The buyer agent then represents the shopper's requirements while the referee provides an independent verification layer.

This multi-agent architecture is one of OpenMarket's most distinctive characteristics.

OpenMarket and Universal Commerce Protocol

OpenMarket is also connected to the developing ecosystem around agentic commerce protocols.

M11 Labs says OpenMarket runs on the Universal Commerce Protocol (UCP), described in the company's launch materials as an agentic-commerce standard co-developed by Shopify and Google, with support from companies including Amazon, Meta, Microsoft, Salesforce, Stripe, Etsy, Target and Wayfair.

The importance of this approach is that AI agents need structured ways to interact with commerce information.

If agents are going to discover products, understand product data and eventually participate in purchasing workflows, they need reliable access to current commerce information.

M11 says OpenMarket can work with live catalog data through this protocol.

How Many Products Are Available?

M11 Labs has described OpenMarket as spanning millions of merchants and billions of products.

The company's launch materials say the platform reaches more than 3 million merchants through open marketplaces including Shopify and covers billions of products.

That scale is significant because the usefulness of an AI marketplace depends heavily on how many real products it can access.

A small product database can limit recommendations.

A marketplace connected to a much broader commerce ecosystem can potentially provide more competitive comparisons.

Does OpenMarket Actually Purchase Products for You?

One important detail is that OpenMarket is not presented as a conventional autonomous checkout system.

The current research preview focuses on the agent competition and recommendation experience. A Product Hunt discussion explicitly notes that the buyer agent cannot purchase anything and that there is no checkout in the current preview.

This means the human remains in control of the final decision.

That is an important distinction because the platform is experimenting with AI-mediated shopping without completely removing the shopper from the process.

Can Seller Agents Negotiate?

Another interesting possibility is negotiation.

M11 representatives have discussed seller agents competing around product quality and claims, and the launch discussion also mentions experiments where seller agents can potentially compete through pricing, discounts or bundles.

If this develops further, shopping could move beyond simple product comparison.

Instead of:

“Which product should I buy?”

the buyer could potentially have agents negotiate:

“Can this seller offer me a better price or bundle?”

That would make the marketplace behave more like a digital negotiation environment.

Who Could Benefit From OpenMarket?

OpenMarket is mainly interesting for consumers who want help researching products, especially when the purchase involves many competing claims.

Potential use cases include:

Electronics

Buyers could compare specifications, battery claims, performance and features across competing devices.

Beauty and Personal Care

Claims around ingredients, testing, performance and certifications could potentially be examined more carefully.

Food Products

Product ingredients, nutritional information and brand claims could be compared.

Fashion

Material, sustainability and product specifications could be evaluated against available evidence.

Expensive Purchases

For higher-value products, having an AI system challenge claims before purchase could be particularly useful.

The broader M11 platform currently identifies electronics, personal care, food, fashion and premium consumer goods among the areas where it works with companies.

What Makes OpenMarket Different?

The strongest differentiator is its multi-agent competitive structure.

Traditional shopping:

Brand → Marketing → Customer

AI recommendation:

AI → Recommendation → Customer

OpenMarket:

Seller Agents → Competition → Buyer Agent → Referee → Customer

That extra layer of competition and verification is what makes the concept unusual.

It attempts to make the reasoning process more visible instead of hiding everything inside one recommendation score.

Limitations of OpenMarket

OpenMarket is promising, but there are several things to keep in mind.

First, it is currently a research preview, so the experience should not be treated as a finished traditional shopping platform. M11 itself describes the product as an experimental public launch with rough edges.

Second, evidence verification is not the same thing as physically testing a product.

An AI system can check available documentation, sources and claims, but that does not necessarily mean it has independently tested the product itself.

Third, the quality of an AI marketplace depends on the quality and availability of the underlying product data.

If a brand provides incomplete information, the AI may have less information to work with.

Finally, consumers should still make their own decisions for important purchases. An AI recommendation can support research, but it should not automatically be treated as an absolute guarantee of product quality.

OpenMarket Pricing

OpenMarket is currently presented as a free research preview in its Product Hunt listing.

Because the platform is still being developed, users should check the official OpenMarket website for the latest access conditions and any future pricing changes rather than assuming the current research-preview model will remain unchanged.

Final Verdict

OpenMarket by M11 Labs is one of the more unusual experiments in AI-powered shopping.

Instead of building another chatbot that simply recommends products, it creates a marketplace where different AI agents can represent competing products, challenge each other's claims and use an independent referee to evaluate evidence.

The idea is simple but powerful:

A product should win because it can prove that it deserves to win—not simply because it has better marketing.

OpenMarket is especially interesting because it represents a possible future where AI agents become active participants in commerce.

The shopper describes what they need. Seller agents compete. A buyer agent protects the user's requirements. A referee checks the claims. The human then remains in control of the final decision.

The technology is still in its research-preview stage, so it is too early to know exactly how successful this model will become. But the underlying concept is worth watching closely.

As AI becomes more involved in product discovery and online purchasing, trust, evidence and verifiability could become just as important as price and advertising.

And that is exactly the problem OpenMarket is trying to solve.

If the future of shopping is going to be powered by AI agents, OpenMarket offers an interesting glimpse of what that future could look like: products competing in an open marketplace where evidence matters more than marketing.