Cordant exits stealth with $8m to close the payments control gap

Cordant has emerged from stealth with an $8 million seed round from Motive Partners, Oak HC/FT and others, pitching itself as a command centre for financial infrastructure.

For chief executive and co-founder Eric Rosenthal, the company is the answer to a control gap he and his three co-founders kept running into at payments firm Rapyd, where a single transaction could touch banks, a processor, a ledger, a compliance tool and a local partner without any one system holding the whole picture.

"To maintain command and control, enormous effort went into managing the complexity," Rosenthal says of his Rapyd years. "Yet inevitably, when something went wrong, teams had to pull information from different systems and counterparties to reconstruct what happened, where the records stopped matching and which controls had been applied."

That problem gets harder as the infrastructure expands. Real-time payments, stablecoins and other new rails add more systems, counterparties and rules to environments that were already fragmented, and AI raises the stakes again. "If an institution wants software or an agent to act on a payment, clear an exception or automate another regulated workflow, it first needs enough context to understand and explain the decision," he says. "We've said from the beginning that context has to come before automation. Regulated institutions will need to explain to regulators and auditors what their agents did and why in real-time."

A neutral view across the seams

Cordant sits within the infrastructure a company already runs. It ingests signals from payment rails, bank accounts, ledgers, and compliance and risk systems, normalises them into a shared event model, and compares what actually happened against internal and external rules documentation to surface contradictions in real time. Payment operations, treasury, compliance, risk and audit teams get a common view across systems that normally show only their own part of a transaction, and the platform can produce an audit-grade record showing how a transaction moved, who acted and which controls were applied.

Just as deliberate is what it does not do. Cordant does not move or hold funds, does not replace core systems, and does not require an institution to centralise its data into a warehouse or lake. "That separation is important because the problem often sits between systems rather than inside one of them," Rosenthal says. "If Cordant became another system of record, a participant in the flow of funds, a data warehouse, or another technical observability platform, we would be adding to the infrastructure we are trying to help institutions understand. The goal is to give teams a neutral view across their existing environment without requiring a rip-and-replace project or expanding who can touch production systems."

What eleven design partners taught them

The company has been built with 11 design partners spanning banking, payments, embedded finance, cross-border payments, stablecoins and digital assets; Bitso and Paxos are both investors and design partners. The use case that surprised the team early on was reconciliation. "It still isn't completely figured out. Not in the industry, and not for lack of tools," Rosenthal says. "We have more compute and better models available than at any point in history, and settlements are still being resolved in spreadsheets, by people deciding which of two counterparties is right."

The team came in assuming most of the difficulty sat inside a customer's own systems. It did not. "Companies have spent a decade getting their own house in order. The exceptions come from the boundary with a counterparty, where nobody has authority to set a standard. So we build for the space between companies," he says. What arrives across that boundary is rarely clean: a settlement file at 4am, a portal export, a statement, four partners describing the same movement of money four different ways. "We build on the premise that no counterparty will change anything for us."

Nor, he argues, can the problem simply be handed to a large language model. "Give an LLM four files and it'll tell you in seconds that they disagree. Ask which is right and it can't, because that depends on how a specific counterparty encodes a reversal, what their cutoff is in practice, and what was knowable at the time. None of that data is in the files or even documented. That's what the eleven partners are teaching us."

The investors, and what comes next

Both lead investors, Rosenthal says, came back to the issue the founders had lived first-hand: financial institutions are adding new rails and systems faster than they can maintain a trusted view across them. Oak HC/FT described the problem as living in the seams between banks, processors, ledgers and compliance systems, while Motive Partners focused on the manual reconciliation that persists even as newer rails such as stablecoins arrive. The backers also pointed to the team, four Rapyd alumni who worked across commercial strategy, product, engineering, sales and partner engineering as that business grew across more than 50 countries and more than 100 partner integrations. The $8 million seed round is being used to bring the platform into production and activate private beta customers.

Asked where the category goes next, Rosenthal is categorical. "Every regulated institution moving money ends up with a command centre for their financial infrastructure. There won't be an alternative," he predicts. He points to industry estimates that around half of internet traffic already comes from machines rather than people, and argues that share only goes one direction. "Money is on the same curve, just behind it. Agents are starting to initiate payments, approve them, settle them. In 10 years the overwhelming majority of transactions will be machine to machine, and no human will look at any of them until something goes wrong."

For a regulated institution, he frames the gap Cordant fills as a supervisory one: being able to prove instantly what moved, when, and on whose instruction. "When software is acting on your behalf at machine speed, being able to prove instantly is the entire basis for being allowed to operate in a more autonomous way. Without certainty, autonomy is just faster exposure with your licence attached."

"What stands in the way is that it doesn't look urgent yet. The settlements clear, the books close, and the cost hides in headcount. That holds right up until the machines are faster than the people checking them, which is sooner than the industry is planning for."