The firm says the new product enables banks to classify and label customer transactions from free text narrative automatically and with a 'high degree of accuracy' in different languages. Fast and accurate classification of transactions is important for many banking use cases, from next best product, cashflow prediction, customer budget advice, peer grouping, customer attrition analysis and many others.

Packaged as an integrated part of the vendor's core bank offering, the new tool is also acceible independently via an API plugin for non-customers.

In a pilot with BlueShore Financial, the large language model was used to provide personalised insights on income and expenses. Within their banking app, customers could view income from various sources such as salary, interest, dividends, and rental income, with sections aligned with Canada’s tax categories to check their intra year tax burden reconcile end-of-year tax reporting. A ‘Year in Review’ feature, with statistics on spending and saving, and comparisons to others in their cohort, provided users with a way to view their financial highlights.

Prema Varadhan, president product and COO, Temenos, comments: "Building on our experience with Explainable AI, Temenos is now harnessing the power of Generative AI as the next step in a strategic AI roadmap, powering the future of banking with ethical, secure, and responsible deployment. We believe that such transparency and explainability will be critical for all AI use cases in banking to help satisfy regulatory requirements.”

Transaction classification is the first Temenos product to leverage Generative AI and the company is exploring further use cases for the technology, such as in chat bots and customer interfaces as well as guiding customer journeys and triage and routing customer queries.