Research firm Evident released a report that explores and captures the emerging best practice among banks for understanding the artificial intelligence opportunity; measuring success; ideating use cases; prioritizing their development and reaching the delivery stage at which outcomes can be measured.
Key takeaways
1. While banks have experimented with AI use cases for years, and yielded significant results to date, the last year has seen a step-change increase in ambition and investment. Banks are generating more ideas for AI use cases than ever before, face more pressure from leadership demanding ROI, and need to keep up with the rapid pace of AI innovation. Working out how to scale up use cases, deliver value, and orchestrate the AI activities across the company has, in many banks, become the mandate for newly established group AI leadership teams.
2. AI teams in leading banks are focused on building capabilities across five core priority areas: Χ MAP current, and potential, AI use cases Χ MEASURE the value of those use cases in terms of outcomes Χ IDEATE the most relevant and addressable use cases for the bank Χ PRIORITISE which uses cases to pursue Χ OPERATIONALISE those use cases to deliver results
3. Map: Use cases are the building blocks of AI delivery. The exact definition of a use case can vary between banks (is it referring to the individual AI model or the collection of models in a sophisticated business proposition, for example). However, AI leaders in banks must establish a consistent and standardized classification and terminology.
4. Measure: Shareholders and senior leaders are increasingly demanding tangible outcomes from AI investments. Banks need a common methodology and process to measure, track and report on the value created by their existing (and future) AI use cases.
5. Ideate: The best use cases are intimately tied to business problems, and ChatGPT has led to the proliferation of ideas of AI use cases like never before. Leading AI teams are investing in initiatives to fuel and harness this bank-wide AI ideation, such as increasing AI literacy; embedding AI teams within the business lines; investing in cross-organization knowledge sharing amongst AI talent; and establishing central use case ideas libraries.
6. Prioritize: There may be few limits to the opportunities offered by AI – but delivery resources are always constrained. Banks need a robust, standardised and aligned process to prioritise AI use cases for delivery. This has to cover ROI, operational capacity, risk and governance issues.
7. Operationalize: Delivering value from AI at scale requires that foundations be well laid. Banks we’ve interviewed are focused on investing in foundational AI tools; delivering on long-term data strategy; establishing external partnerships in priority areas where they lack in-house expertise; and ensuring that model validation frameworks are fit-for-purpose, encompassing Generative AI.
8. The race for AI outcomes is only accelerating, and best practices across the sector are beginning to emerge. To that end, we have built an initial list of KPIs that banks can use to assess their progress against the five capability areas explored in this report. I