WHAT IS
AGENTIC BANKING?
ARTICLE · 002Quaestrum
DBDavid BörgerFounder

July 27, 2026

What is agentic banking? A 2026 definition

Agentic banking is banking in which AI agents authenticate, retrieve context and execute financial tasks such as payments, card management and reconciliation, while humans define the policies those agents must obey. The difference to a chatbot in a banking app is execution: an agentic system does not tell you how to pay an invoice, it pays the invoice, inside a mandate a human has authorised.

The three layers of an agentic banking stack

Every serious implementation splits into three layers. At the bottom sit licensed rails: accounts, card issuing, payment execution and KYB, provided by a bank or a regulated infrastructure partner. In the middle sits a control layer: deterministic policies, budgets, approval thresholds and an audit trail. On top sit the agents themselves, connecting through interfaces like the Model Context Protocol, an API or a CLI. The control layer is the new piece. Rails existed before, agents existed before, but the layer that translates human intent into machine enforceable limits did not.

Who is building it

In the United States, Meow launched banking for AI agents in April 2026 on top of its existing business banking platform, and Catena Labs, founded by Circle co founder Sean Neville, raised 48 million dollars and applied for a national trust bank charter to build an AI native financial institution. On the network side, Visa took agent initiated payments live in Europe in July 2026 with more than 30 issuers, and Mastercard ran its first live European agent payment with Santander in March 2026. In Europe, Quaestrum is building the business account for AI agents on licensed European banking rails, with the control layer as the product.

Venture funding for US agent banking, mid 2026 (USD millions)

Catena Labs$48M
Meow$27M
Source: Company announcements, Crunchbase, PitchBook, as of July 2026

Why it matters now

According to Gartner, only 17 percent of organisations had deployed AI agents by 2026, but more than 60 percent expect to within two years, the steepest adoption curve of any emerging technology Gartner measures. Every one of those deployments eventually hits the same wall: the agent needs to spend money, and pasting a corporate card into a prompt is not a spending policy. Agentic banking is the answer to that wall, and the category is being defined right now.

17%

of organisations had deployed AI agents by 2026

60%+

expect to deploy within the next two years

Source: Gartner, 2026 CIO and Technology Executive Survey