ANNA Money has completed a test in which a large language model (LLM) agent successfully completed the entire business registration process through the company's web interface, including full authorisation, identity verification, and payment.
The process of registering a business in the UK typically involves submitting detailed information, verifying identities, and handling online payments, which can be burdensome for new business owners. ANNA Money utilised an LLM agent acting as if it were a human customer to carry out these steps autonomously.
In the test, the AI agent was able to collect and provide all required details, navigate secure login processes, complete stringent identity checks, manage payment, and submit the application to Companies House. Following the completion of the process, a new business account was opened automatically, with a virtual card issued and available for immediate use.
"Frankly, the fact it worked first-time was extraordinary. LLM agents can struggle with something as common as checking in for a flight, yet here we saw it handle one of the most complicated and regulated sign-up processes in the UK - end-to-end, payment included. That's a real milestone for agentic AI in financial services," says Boris Diakonov, co-founder and co-CEO at ANNA.
According to ANNA Money, while the successful demonstration was based on its current systems, steps are already being taken to optimise the interface for interaction with LLM agents and similar AI systems. The company's roadmap includes the development of a dedicated "robot-ready" integration. This enhancement is intended to make it possible for AI-powered assistants, potentially embedded within various platforms, apps, or devices, to register companies and open business accounts on behalf of users within minutes.
Boris Diakonov further commented, "By being first to prove this capability, ANNA is positioning itself at the forefront of the AI-powered economy, where autonomous software doesn't just assist people, it acts on their behalf to get things done. It's a breakthrough that demonstrates ANNA's leadership in financial innovation and its readiness for the next generation of digital business services."
Process details
The demonstration involved inputting all mandatory data for a new UK business, navigating through regulatory checks, and conducting a required payment online. The LLM agent handled each step autonomously, communicating with the interface and completing the account opening without human input. The process concluded with immediate access to a virtual payment card for the new business account.
Industry context
Financial onboarding, especially for new company incorporation, is generally considered a complex and highly regulated area, often presenting hurdles for automation due to security and compliance requirements. The ability of LLM agents not only to support users in these processes but to enact them start to finish adds momentum to the movement towards agentic AI, in which software executes tasks independently on behalf of users.
ANNA Money's plans to further develop interfaces for autonomous agents suggest a future in which AIs could be delegated more operational responsibilities. This shift could streamline administrative tasks for small business owners in the UK and beyond.
Future developments
The company has stated its intention to expand the functionality of its business account and tax management offerings, including the introduction of expense management and credit cards for UK customers. It is also expanding its presence in international markets, having recently entered Australia.
The LLM-driven registration process has implications for the broader adoption of agentic AI in financial services, where regulatory demands traditionally slow automation efforts. ANNA Money's demonstration provides a public instance of an end-to-end complex onboarding operation, completed by a non-human agent within legal frameworks.
