Baselayer raises $35m to trace an agent's payment back to the business that authorized it

The person this lands on is the analyst who sees a payment initiated by software on behalf of a customer and has to decide, quickly, whether to release it. When an agent starts a transaction, according to SiliconANGLE's account of the problem, banks and merchants cannot tell which agent it is, who it is acting for, or whether it should be trusted ^[6]. The rails underneath were built to identify people and registered businesses ^[20].

What Baselayer is selling is narrower than "trust". The Agentic Identity Suite, built on what the company calls "Know Your Agent", traces an agent's actions cryptographically back to the platform that deployed it and the specific business that authorized it. The chain of liability it produces is auditable, and there is no shared secret in it that can be stolen ^[7]. Chief executive Jonathan Awad said the company already helps "one in five U.S. financial institutions answer, 'Can I trust this business?' Now we're building the infrastructure they need to answer, 'Can I trust this agent, who does it represent, and what is it allowed to do?'" ^[8]. The first two of those are lookups. The third is a policy your own customer has to have written down. A bank can confirm that an agent was deployed by a known platform for a known company and still not know whether that company meant to let it move that much money.

The number carrying the market case is Stripe's. In June, Stripe said more than 70% of the requests hitting its API came from AI agents ^[4]. That counts requests, not dollars, and a coding agent calling an endpoint looks the same in that total as an agent paying an invoice. Visa, Mastercard, American Express and Shopify have all shipped agentic commerce protocols this year ^[5], so merchants are taking these orders regardless of how the request share is read.

Baselayer's better asset is the customer list it already has. Its business verification tools are used by more than 2,300 financial institutions, including one in five US banks ^[10]. The company says the platform has reviewed more than 10 million applications and prevented over $1 billion in fraudulent transactions ^[11]. Divide one by the other and you get roughly $100 of prevented fraud per application reviewed ^[12]. That is a lot of small decisions.

Baselayer calls the suite the first interoperable trust layer and fraud prevention platform for the agentic economy ^[3]. It is also working on agent identity inside the FIDO Alliance Authentication Working Group, the Legal Context Protocol and the x402 Identity Working Group, alongside Google, Cloudflare, Visa and Mastercard ^[16]. If those groups settle the primitive in a spec, what Baselayer still owns is distribution into 2,300 institutions.

The SiliconANGLE report does not include a price or an availability date for the suite ^[19]. For anyone who has to roll this out, the useful sort is two questions per agent transaction: can you identify the agent and its owner, and can you state, from your own records, what that owner permitted it to do. The cell where the answer is yes then no is where a vendor helps least.