Why We Invested in Skalor: The Missing Control Layer for Agents and Agentic Payments

Draper Dragon has invested in Skalor, which is building the control layer for AI agents - in effect, an "agentic clearing house" for agent actions and decisions.

Companies are giving AI agents the power to take actions and, critically, to spend money faster than they are building a way to control them. However, the pieces for this control are already in place. We know how to prove who an agent is. We have rails for it to pay on. We have directories for it to find information. What is missing is the part that decides whether a payment or action should happen at all. An agent that has been tricked, or has simply drifted from what it was told to do, can do a lot of good or a lot of destruction (including spending real money at machine speed – with the finance team finding out the next morning).

Skalor sits in front of anything an agent tries to do, whether that is a payment, a transfer or a decision, and checks it against the rules before it goes through. Companies set limits: how much an agent can spend in a day, per supplier, per transaction. They set a list of who it is allowed to pay, and everything else is refused. If an agent starts behaving strangely, Skalor freezes it and sends it to a person to review. Every decision it makes is recorded in a form that cannot be altered afterwards, which is what a fraud team and a regulator will each ask to see. All of it happens in a few thousandths of a second, which matters more than it sounds: a safety check that slows the business down is a safety check that gets turned off.

Two things convinced us to invest. Skalor never holds client funds. The keys stay with the customer, and Skalor only ever says yes or no. That is the difference between something a bank's risk committee will approve and something it will not. The second is that Skalor works across whichever rails a company ends up using, rather than betting on one winning. Nobody knows yet which rails win. Every company will need brakes either way.

Backing Skalor is a continuation of how we have invested for a decade – in the infrastructure that makes new financial plumbing safe enough for serious institutions to use. Our portfolio already covers much of the payments layer that agent commerce will run on, and some of the first companies building agents that transact on their own. Skalor is the next piece of the puzzle.

"Every enterprise deploying AI agents with real treasury access is going to need policy enforcement before those agents can take action and move money at scale," said Samiz Bayan, partner at Draper Dragon. "Skalor's non-custodial approach means enterprises never have to hand over their treasury or their executive controls to get that protection."

Learn more about Skalor here.