Rezolve Ai, and the question of trust

The role

Dan founded Rezolve in 2016, after Powa, with a narrow initial idea: that any piece of advertising or media should be instantly shoppable from a phone. The company has since widened into what it now calls infrastructure for the agentic economy. He has led it throughout, first as a private company and, since its Nasdaq debut on 16 August 2024 under the ticker RZLV, as a public one. His formal title is Founder, Chairman and Chief Executive Officer.

In practice the job has three parts. He sets the product thesis, described below. He has led the company's partnership and acquisition programme, including the relationships with Microsoft and Google and the purchase of the Subsquid data infrastructure business in October 2025. And he is the company's public voice, on earnings calls, at conferences such as Web Summit and NRF, and on the daily feed of posts collected on the Thinking & Writing page.

The company, briefly

Rezolve Ai sells software to retailers, brands and financial institutions. Its product family is grouped under the Brain name: Brain Commerce for search and recommendation, Brain Checkout for policy-controlled purchasing by people or agents, and brainpowa, the company's own commerce-tuned language model. Around these sit TraceWare, which records how an AI system reached a decision so that the result can be audited, and a payments layer, RezolvePay, designed for both card and stablecoin settlement. Since 2024 the company has added several businesses by acquisition, among them GroupBy, ViSenze, Crownpeak, Smartpay, Truther, Subsquid and the Reward loyalty network.

The company reported unaudited revenue of $130.8 million for the first half of 2026, more than twenty times the same period a year earlier, and more than 1,640 enterprise customers at 30 June 2026. Its guidance for full-year 2026 revenue is around $360 million. These are the company's figures; the audited record is in its filings with the US Securities and Exchange Commission.

The argument

A general-purpose model can talk. Very few can actually sell, and fewer still can be trusted to execute a transaction without hallucinating a price, a product or a promise.Dan Wagner, LinkedIn, 2026

Dan's position on AI in commerce has been consistent since long before it was fashionable, and it is not the optimistic one. He has argued since at least 2016 that large language models are probabilistic by design and will therefore always carry a margin of error, and that in a shop that margin is not an awkward answer but a refund, a chargeback and a damaged brand. No amount of prompting or "guardrails", in his view, removes the flaw; it has to be engineered around.

The conclusion he draws is that the interesting layer of agentic commerce is not the agent that can pay, which the large platforms and payment networks are building quickly, but the layer that makes an agent's purchase trustworthy: discovery grounded in the merchant's real catalogue and prices, execution governed by the merchant's rules, and an evidence trail that shows what was done and why. That is what Rezolve's products are built to provide, and it is why he describes the company's work as infrastructure rather than as a shopping assistant.

He is equally direct about payments. He has welcomed the Bank of England's more open stance on stablecoins and argues that regulated stablecoin settlement, invisible to the merchant and the customer, will do to card economics what real-time payment schemes have done in India and Brazil. Rezolve's collaboration with Tether, announced in December 2024, and its plan, announced in February 2025, to hold bitcoin as a treasury asset both follow from that view.

The record of these positions, in his own words and with their dates, is on the Thinking & Writing page. Where they have drawn scrutiny, from analysts, short-sellers or the press, the Press & Archive collects the reporting.