Starling unveils an AI banking assistant that truly performs tasks.

      The UK challenger bank is launching Starling Assistant for personal account holders today, promoting it as the UK's first agentic AI financial assistant. It can assist in setting savings goals, managing bill payments, and even quizzing users on their own spending—all through a voice or text command.

      Starling Bank has been systematically preparing for this development for nearly a year. In June 2025, it became the first UK bank to implement a natural language AI interface for customer spending data through Spending Intelligence. In October, it introduced Scam Intelligence, allowing customers to upload images of marketplace listings or suspicious messages to receive a fraud-risk assessment. Both services operated on Google Gemini via Google Cloud and were marketed as UK firsts. On Friday, the bank unveiled the next phase: Starling Assistant, which it claims to be the UK’s first agentic AI financial assistant.

      The distinction is significant, at least theoretically. Spending Intelligence and Scam Intelligence are tools for reading and responding—they analyze data and provide outputs. In contrast, Starling Assistant is designed to take action by responding to voice or natural language prompts and performing banking tasks directly for the customer. It also serves as the overarching platform for the previous tools, consolidating them into a single conversational interface.

      Its capabilities are more focused than most announcements regarding agentic banking typically promise. For instance, a customer planning a holiday can request to save £500 for a trip to Paris in July and ask the assistant to create a monthly savings plan and set up automatic transfers to a designated Space. Someone wanting to reorganize their finances on payday can direct it to establish Spaces for groceries, bills, travel, and dining out, specifying how much to allocate to each. Additionally, the press release notes that the assistant can answer inquiries about direct debits and outstanding bills, analyze transaction histories with particular payees, and even generate a quiz where the assistant asks the customer to identify their top merchant from the last month or their highest spending category.

      It is important to mention one caveat: voice prompts are activated by the user’s mobile keyboard rather than through any built-in voice recognition technology within the assistant itself. While the difference is minor in practice, it is noteworthy for those expecting a completely hands-free experience.

      The assistant also incorporates a clear welfare aspect. Customers with vulnerabilities or accessibility requirements can request specialized support without needing to speak with a human agent: it can assist hard-of-hearing users in setting up Starling’s sign language service, guide individuals through implementing gambling blocks, and direct those facing financial hardship to appropriate resources.

      Harriet Rees, Starling’s group chief information officer, described this launch as the result of eight years of AI development at the bank. Raman Bhatia, the group chief executive, referred to agentic AI as “the next step in banking.”

      Starling Assistant is now available for personal current account customers, with a rollout for business and joint accounts anticipated soon. Participation is opt-in, and Starling assures that all data remains within its Google Cloud environment and is not used for training the underlying models. Graham Drury, director of FSI at Google Cloud UK and Ireland, emphasized that this shift represents a movement away from navigating intricate app menus toward simply having a conversation about one’s finances.

      The backdrop of this launch is somewhat complex. In October 2024, the Financial Conduct Authority fined the bank £29 million for anti-money laundering and sanctions screening failures from 2017 to 2023, a penalty reduced from £41 million following Starling’s cooperation with the inquiry.

      Establishing a reputation as the UK’s most ambitious AI-driven bank necessitates a solid compliance framework. The welfare features integrated into Starling Assistant, the opt-in approach, and the explicit pledge not to employ customer data for model training align with a bank that has dedicated the past year to rebuilding regulatory credibility alongside its product development roadmap.

      In the broader neobank sector, the competition to implement agentic AI is intensifying. Revolut has indicated it is exploring AI agents but has yet to launch a similar product in the UK. Bunq introduced an AI assistant in 2024, while Klarna has widely utilized AI in customer service. Currently, Starling is asserting that it is the one defining the landscape in UK retail banking.