Starling Bank Launches UK’s First Agentic AI Banking Assistant | TNW
Starling Bank is expanding its foray into artificial intelligence with the launch of Starling Assistant, billed as the UK’s first agentic AI financial assistant. Available to personal account holders starting today, the tool moves beyond simply analyzing financial data – a feature Starling pioneered in June 2025 with ‘Spending Intelligence’ – to actively managing finances based on user prompts. Users can now instruct the assistant to set savings goals, organize bill payments, and even receive personalized quizzes about their spending habits, all through voice or text commands.
From Data Analysis to Actionable Insights
Starling’s initial step into AI-powered finance, Spending Intelligence, allowed customers to ask natural language questions about their spending, such as “How much did I spend on groceries last week?” and receive instant analysis. This feature, built using Google Gemini on Google Cloud, marked the first time a UK bank offered such direct interaction with spending data. The bank followed this with Scam Intelligence, which assesses the risk of online marketplace listings and suspicious messages. Starling Assistant represents a significant evolution, consolidating these earlier tools under a single conversational interface and adding the ability to act on user requests.
The key distinction, as Starling emphasizes, is the “agentic” nature of the new assistant. While Spending Intelligence and Scam Intelligence provide analysis, Starling Assistant is designed to execute tasks. For example, a user planning a trip to Paris can ask the assistant to calculate a monthly savings schedule and automatically transfer funds to a dedicated ‘Space’ – Starling’s internal budgeting feature. Similarly, users can instruct the assistant to create separate Spaces for different spending categories (groceries, bills, travel) on payday, specifying the amount allocated to each.
Beyond Basic Task Management: A Focus on Financial Wellbeing
The functionality extends beyond simple budgeting. Starling Assistant can answer questions about direct debits and outstanding bills, analyze spending with specific merchants, and even generate a quiz to test a user’s awareness of their own spending patterns – for instance, asking them to guess their top merchant from the previous month. This gamified approach aims to encourage greater financial awareness.
Notably, Starling has integrated accessibility and vulnerability support into the assistant. Customers with hearing impairments can use the assistant to set up Starling’s sign language service, while those struggling with gambling addiction can easily activate gambling blocks. The assistant can too direct users in financial distress to specialist resources. This focus on wellbeing is a deliberate design choice, reflecting a broader commitment to responsible AI implementation.
Technical Implementation and Data Privacy
Starling Assistant leverages Google’s Gemini models on Google Cloud, with all customer queries analyzed securely within Starling’s infrastructure. The bank emphasizes that the feature is opt-in, and user data is not used for training the underlying AI models. This commitment to data privacy is particularly essential given the sensitive nature of financial information.
One minor limitation is that voice prompts are currently enabled through the user’s mobile keyboard, rather than native voice recognition built into the assistant itself. While functional, this means the experience isn’t fully hands-free.
Navigating Regulatory Scrutiny and Competitive Landscape
The launch of Starling Assistant comes after a period of regulatory scrutiny for the bank. In October 2024, the Financial Conduct Authority (FCA) fined Starling £29 million for anti-money laundering and sanctions screening failures between 2017 and 2023. The fine was reduced after Starling cooperated with the investigation.
This context underscores the importance of building trust and demonstrating a robust compliance infrastructure alongside product innovation. The opt-in model for Starling Assistant, the commitment to data privacy, and the inclusion of welfare features all appear to be deliberate steps towards rebuilding regulatory credibility.
The Broader Trend of Agentic AI in Fintech
Starling isn’t alone in exploring the potential of agentic AI in banking. Revolut has signaled its intention to develop similar AI agents, though a UK launch date remains unconfirmed. Bunq launched an AI assistant in 2024, and Klarna has extensively deployed AI across its customer service operations.
Starling, however, is positioning itself as a leader in the UK retail banking space, arguing that its approach – combining proactive task management with a strong emphasis on data privacy and financial wellbeing – sets it apart. Harriet Rees, Starling’s group chief information officer, described the launch as the culmination of eight years of AI development at the bank, while group chief executive Raman Bhatia called agentic AI “the next step in banking.”
Starling Assistant is currently available to personal current account customers, with plans to extend access to business and joint accounts in the future. The bank’s success will likely depend on its ability to demonstrate the practical benefits of agentic AI while maintaining the highest standards of security and regulatory compliance. The next phase will involve monitoring user adoption, gathering feedback, and iteratively improving the assistant’s capabilities based on real-world usage patterns.