AI is already making decisions for banks. 2% of AI use cases involve completely unsupervised autonomous decisions, but 55% involve some level of autonomous decision-making (Bank of England). It’s a safe assumption that those numbers will rise.
The question is, what will that look like?
AI’s impact on strategic decisions for banks
The global market for predictive analytics in banking is forecast to be worth $745bn by 2030. That is hardly a surprise — which bank wouldn’t want a rich and data-driven forecast of market and customer behaviour?
Far from making analysts redundant, it will free them to focus less on generating report and insights, and more on guiding AI’s work and refining results. Then, they will also have more time to use those insights to make strategic recommendations. Recommendations like:
- Refining products and offerings
- New products and services
- Operational efficiencies
- Risk analysis
How AI will enrich banks’ commercial decisions
AI will help banks make better commercial decisions, removing inefficiencies on the way to making those decisions, and providing better data quality and quantities to inform them.
Take loan applications for example. Automation can easily handle the more basic assessments, meaning more qualified applications can reach the next stage quicker. That means:
- Humans are freed to do more detailed and in-depth assessment that require discretion and lateral thinking
- The bank can accept more applications
- Customers are less likely to choose competitors who can accept applications faster
Another example would be Anti-Money Laundering (AML). If AI can perform better analysis, quicker, then the bank can reject more non-compliant applications sooner, and reduce the number of false positives that cost accounts and alienate prospective clients.
Ensuring your AI models are capable of making and enabling better decisions
Two of the pillars of AI ethics are ‘explainability’ and ‘accountability’, and those must be guiding principles for your AI-driven decision-making.
Explainability: You need to be able to make you AI’s decisions understandable. If challenged, you must be able to present the reason AI did what it did.
Accountability: Even when delegating a decision to AI, the AI may have made the choice in practice, but ethically it was your choice.
To operate in that ethical framework, and generally to ensure that AI is an asset and not a liability, you must first assess and protect the quality of your data.
Data is what trains and fuels your AI, and if it has unreliable, or inaccurate data, then it will be full of unwanted bias and operate with fundamental misconceptions of reality. Then its decisions will be worse than worthless. For any sector, but particularly for Financial Services, Data Governance is a foundational step towards AI decision-making.
When you are confident in your underlying data, you need to be just a certain about your AI model itself. An audit your AI artefacts, will either offer the peace of mind that comes with knowing your tools are safe and effective, or a clear and comprehensive list of work, so that you can take action immediately.
Plotting your AI roadmap
How do you seize the AI opportunity and remain competitive, while avoiding the potentially costly missteps?
Download your guide to AI in Financial Services and:
- Realise the potential for AI in your organisation, while appreciating and mitigating the risks
- Navigate the current and evolving regulatory frameworks and legal obligations around AI in Financial Services
- Understand the future of work in Financial Services following the advance and adoption of AI
Your guide to Artificial Intelligence: Safety, compliance, regulation, and the commercial impact