Trust in e-commerce is at an all time low.

For years, the biggest marketing budget decided what got sold. Better ads and more visibility meant more sales. Of course, product quality mattered, but it mattered less than distribution. A mediocre product with a large budget could easily outsell an excellent product with a small one, and it usually did.

But the buyer has changed, and what worked before no longer does.

Today, the first reader of a product page is usually not a person. It is an AI agent, answering a question, comparing options, or completing a purchase on someone's behalf. And AI does not respond to confident copy. It trusts only what it can verify. So when an agent reads “best entry-level road bike” but finds no reviews, awards, or tests to back it up, it has no reason to trust it. Sure, a brand can say their product is the best on the market, but without outside proof, the agent won’t trust it.

We call this the zero-trust internet. Saying something is true is no longer enough. You have to prove it.

In fact, the scale of this gap surprised even us. In our own benchmark this year, we found that 88% of product claims had no proof behind them (M11 Benchmark, 2026).

The interesting thing is that, in most cases, the product is as good as the brand says, but there’s no proof to back it up.

A garment may be handmade, but if the website doesn’t explain the process behind it, why would an AI trust it? The same goes for a heavy-duty drawer that doesn’t state how much weight it can hold or a safety certification with no test report or lab listed.

And it’s not just AI. The regulations are moving in the same direction. EU substantiation requirements on green claims apply from September 2026, and FTC substantiation standards already treat an unsupported claim as a liability.

The regulatory reader and the AI reader now want the same thing: the evidence.

Introducing, M11 Agents

M11 agents keep brands both accurate and competitive by verifying their claims and detecting threats from competitors and the market. The work covers four key areas.

1. Presence catches the data sources that hurt a brand's identity, the places where AI simply does not surface a product that should rank. 

2. Accuracy fact-checks every claim the brand makes and binds each one to its evidence, or flags the gap. 

3. Completion fixes the data gaps in the product record itself, the missing dimensions, materials, and certifications that agents need to recommend with confidence. 

4. Competition tracks what competitors are doing, from a sitewide discount that signals margin distress to a stockout that opens an acquisition window, and reads what each move means for your brand.

Our agents don’t stop at pointing out the problem, they help fix it. Instead of handing you another report, they produce something you can actually use: a corrected origin statement, a better-supported claim, or a data update based on the maker’s own records. 

We suggest the fix, you approve it, and our agents put it into action. 

The M11 Commercial Score tracks progress, showing how much things have improved and what drove it.

We really do care

We are building M11 Labs because the alternative is worse.

If proof doesn't dictate what agents recommend, spend will. The zero-trust internet will become a flood of low-quality content: fake reviews, made-up authority, and the same budget race, just with AI making the calls.

Right now, there is a narrow window where the rules of this new internet are being written. We want evidence to be the standard.

But our drive to build this is also deeply personal. Throughout our careers, we have watched brands make genuinely exceptional products, only to get crushed by inferior competitors with massive ad budgets. Watching a well-made product lose to a well-funded one never stops being frustrating.

Now, for the first time, the infrastructure of commerce is being rebuilt. We finally have a chance to reward the maker over the marketer.

And that is a mission worth building for.

We are building a world-class team

Behind M11 Labs is a world-class team working across London and San Francisco, building the new standard for commerce.

Ankur Modi, our CEO and co-founder, has spent his career building commerce AI at the largest scale it exists. At Amazon and Meta, he engineered the systems that dictate what gets surfaced and what gets sold, driving billions in incremental revenue. Having scaled a company through a NASDAQ IPO and advised the UK government on AI policy, Ankur knows the internal logic of commerce platforms better than anyone. M11 is his answer to what happens when that logic is handed over to AI agents.

David Mataciunas, our CTO and co-founder, tackles the same problem from the AI research side. With experience at IBM Research and Cohere Labs, and as chair of the Lithuanian AI Association, David understands the architectural limits of language models. His research, published at NeurIPS, focuses on exactly what LLMs cannot reliably do. He knows firsthand that AI cannot be trusted to "know" the truth, which is exactly why he built M11’s agent swarm to force models to rely on verified, structured evidence instead.

––––

Does  AI trust your brand? Ask M11 agents now