The M11 Agents That Help Good Products Win

When an agent reads my product, what does it find? This is the first question every brand should ask to understand how AI sees its products.

Data from the Age of Online Trust Report notes that it takes 4 positive experiences to build consumer trust, but only 2 bad experiences to destroy it completely.

What does this mean for your brand? If an agent can’t read, verify, and trust your products, you may not appear in its recommendations at all.

At M11 Labs, we built agents that help brands earn trust from AI agents. Our agents verify what brands say about their products and monitor competitors and the market so brands can spot problems, stay competitive, and keep their information accurate.

The work covers four key areas.

Accuracy: can AI trust what you say?

Accuracy fact-checks every claim the brand makes and binds each one to its evidence or flags where the evidence is missing. In the M11 Labs Benchmark this year, 88% of product claims had no proof behind them (M11 Labs Benchmark, 2026). Most claims are far from being lies. They are true statements without evidence, and to an agent the two look identical.

The stakes are rising on the regulatory side too: ECGT enforcement begins on 27 September 2026, with fines of up to 4% of annual turnover for unsubstantiated environmental claims, and member state authorities have already fined brands between €1 million and €25 million ahead of it. 

M11's Accuracy agents extract every claim, hunt every evidence tier, grade the fit, and draft the fix: a claim rewritten so it survives both an agent's and regulator's check.

Completion: is your product record whole?

Completion fixes the data gaps in the product record itself. Think of a drawer with dimensions listed but load capacity missing or a premium fabric with no composition mentioned on its product page. Agents compare options against constraints, and an unanswered constraint is a lost sale as the agent skips the product without a human ever seeing it.

M11's Completion agents find the live page per SKU, extract every data block, score 27 structured fields, and build the spec from the maker's own data. Across live brand deployments, our agents take data completion to 90%+ against a baseline of roughly 36% and produce 250%+ more agent-readable data than standard.

Presence: does AI find you at all?

Presence catches the data sources that hurt a brand's identity, the places where AI does not surface a product that should rank. Think of a non-alcoholic aperitif missing from Amazon results for its own category or a brand ranked behind a weaker competitor because the rival's data is easier to read. 

M11's Presence agents ask what real buyers ask, query every AI surface, extract the mentions and citations, and then publish the data AI needs to rank you. Given how well the agent channel is converting, being absent from it is no longer just a discoverability problem. It can directly affect your revenue.

Competition: what is moving against you?

Competition tracks threats from competitors and the market and reads what each move means. A competitor cutting prices by 50% may signal that they are under pressure. A competitor running out of stock at a major retailer may create an opportunity to win those customers.

These changes are happening faster, too. According to Adobe Commerce (2025), competitor catalogue changes were 3x higher than the previous year. No team can realistically keep up with that pace manually.

M11’s Competition agents track what competitors are doing and flag the changes that matter to your brand. After that, the agents suggest what to do next.

How do the four parts come together?

A competitor’s unsupported claim can reveal both a competition threat and an accuracy problem. A missing piece of product information can explain why a product is not showing up in AI results. We built one number for this, the M11 Commercial Score. Our agents feed all four signals into it, and it shows where a brand stands and what needs attention.

Behind the score are 16 specialised agents, 240+ sub-agents per scan, and 11 frontier models working together continuously. Each agent goes beyond finding a problem. It drafts the fix.

M11 agents propose the change, the brand approves it, and the agents deploy it. Pulse shows the score, Command manages the signals, Proof builds the evidence, and Graph connects each claim to the evidence behind it.

Why can't a platform or a chatbot just do this?

#1 No single sales platform sees across every channel a brand sells through, and they compete with each other; our agents are structurally neutral. 

#2 Between 40% and 60% of the proof M11 agents use come from private sources no scraper or foundation model can reach: ERP and lab reports, certification registries, clinical and compliance records. 

#3 Knowing which claim needs which proof from which source is standards-layer expertise that compounds with every brand and category. "Carbon neutral" needs LCA data. "Clinically proven" needs an RCT citation. A chatbot does not know that, and a commerce platform has no reason to build it.

What does M11 deliver?

So far, M11 agents have delivered 250%+ more agent-readable data than standard, 90%+ data completion compared with a roughly 36% baseline, and the first signals within 48 hours of connecting.

These numbers reflect real problems we have helped brands solve. A personal care brand found dozens of clinical claims with no sources behind them. A CPG company found SKU-level data gaps its team had missed for a year. An electronics brand found misleading competitor information affecting its AI recommendations. M11 agents helped each brand find the problem and fix it.

That’s what we built M11 Labs for: finding what’s missing, helping brands fix it, and making sure good products have a better chance of being chosen.

–––––

Does  AI trust your brand?

Ask M11 agents now.