Starling Assistant Is Customised ~ Is this the Future of Money Management?

August 2026

Fintech & Payments

Thomas Wilson

Senior Research Analyst

Starling Bank is turning its AI assistant into a platform. With the launch of weekly ‘Smart Tools’ inside the Starling Assistant, the bank is not just adding more prebuilt features; it is opening a path for customers to request, shape and, ultimately, create their own AI‑powered banking tools. For British citizens, this shift could redefine everyday money management, and if executed well, materially improve fraud prevention.

What Is Starling Actually Launching?

Starling’s Smart Tools are AI ‘skills’ that plug into its agentic assistant, the Starling Assistant, which already understands natural language and can act on a customer’s behalf. The initial set includes capabilities such as personalised savings plans, automated VAT ring‑fencing for businesses, and guidance on scams. Crucially, the bank says it will release new tools every week, and is crowdsourcing ideas from customers, with a longer‑term ambition to allow users to build and submit custom tools. In effect, Starling is moving from a fixed menu of AI features to a more open, extensible model where the assistant can be tailored to individual needs.

This mirrors how Starling’s own engineers use AI internally. They encode recurring tasks as skills, so coding agents can execute them reliably at scale.

How Will this Affect Money Management?

For UK customers, the most immediate impact is on budgeting and saving. Tools such as the Rainy Day Saver analyse income, direct debits, and spending patterns to estimate how much a customer can realistically save, then automate transfers into a dedicated Space. Over time, as more Smart Tools arrive and customisation deepens, users could assemble a personal money stack inside one app: cashflow forecasting, tax provisioning, subscription auditing, travel budgeting, and more.

This matters because British households face volatile energy bills, mortgage rate resets, and a high cost of living. An assistant that can both reason about a customer’s finances and act - taking action such as moving money, creating rules, and flagging anomalies - shifts money management from passive dashboards to active, automated control. For self‑employed people and small businesses, automated VAT ring‑fencing and Making Tax Digital (MTD) readiness reduce administrative burden and the risk of cashflow surprises.

The question for the wider market is obvious. If one bank proves that user‑configurable AI tools improve outcomes, will rivals follow? If so, the default expectation for retail banking in the UK could shift from static features to personalised, evolving AI workflows.

A New Tool for Fraud Prevention

Fraud is where Starling’s approach could have the clearest national impact. The Fraud Control Smart Tool is designed to help customers turn on the right protection controls for their profile, including Snatch Theft Detector, Scam Intelligence, and call status indicators. Scam Intelligence, in particular, uses multi‑modal AI to assess images. For example, screenshots of messages or payment requests, and surface red flags.

By surfacing these controls inside a conversational assistant, Starling lowers the friction to enable stronger protection. Instead of hunting through settings, a customer can ask the assistant to ‘turn on anti‑scam protections’ or ‘show me how to spot invoice fraud’, and the tool will guide them through configuration and education. Over time, as custom tools emerge, we could see SMEs building sector‑specific fraud workflows, such as automated checks for supplier invoice changes, or rules that flag unusual international payments.

This is significant for UK fraud trends. Authorised push payment (APP) fraud losses have remained stubbornly high, with consumers and small businesses particularly exposed to social engineering and invoice fraud. If AI assistants can make advanced controls easier to understand and switch on, and can embed contextual guidance at the moment of decision, the industry may finally see a step change in prevention, rather than just post‑event reimbursement.

However, it can be argued that many customers are unlikely to actively configure advanced fraud controls for themselves. Fraud features are inherently invisible until they work, or worse, until they block something important. For the average user, the default assumption is that the bank already protects them, and being asked to choose and tune different protections could feel like the bank is offloading responsibility onto the user, rather than adding value.

Wider Implications

For everyday users, the promise is simpler: money management that adapts to your life, not the other way round. For small businesses, it is less admin and more real‑time financial control. For fraud, it is a shift from hoping the customer reads the warnings to embedding the right controls and guidance in the flow. 

The bigger question for the UK market is whether Starling’s experiment will force a reset on personalisation and control across retail banking. If custom AI tools prove their value, the next wave of competition may not be about interest rates or cashback, but about who enables customers to design the most effective, safest, AI‑driven money workflows.