How Can Financial Institutions Ground LLMs for Real Operational Use?
June 1, 2026An illustration of how probabilistic AI can be grounded inside deterministic financial operations
Following on from our December 2024 research into the use of artificial intelligence and data analytics within credit risk management processes at financial institutions, we undertook a complementary analysis of the leading solution vendors in this field.
See here for our summary article of the initial use case research: https://distinctiveinsights.ai/ai-credit-risk/
As with the initial research, this follow up analysis was based upon data analysed for the period 1st January 2023 to 1st November 2024.
This article aims to provide an easily digestible summary of some key outcomes from this research. The full research is available for subscribers to download on our website.
We highlight Nova Credit as a key vendor which is developing innovative credit solutions in partnership with the global banking market segment. Its solutions enable new-to-country populations to access and share their credit history, extending financial access and inclusion to newcomer groups without credit history within their new country. Global banks that Nova Credit has announced partnerships with over this period include HSBC, Royal Bank of Canada and Scotiabank.
This trend within global banks tracks developments within the U.S Credit Union sector where access to credit is being broadened by solutions connecting a significantly wider universe of ‘alternative’ data than traditionally used in credit assessment methods. Leveraging artificial intelligence-based solutions is facilitating the connectivity of an extensive set of new data to a credit applicant’s profile, overcoming the limitations of the traditional credit file. A prominent example of a leading vendor in this process is Scienaptic which utilises an array of alternative data including telecoms bills, utility bills, rental payments data, and a large amount of additional content to create a more complete view of an applicant’s creditworthiness. Scienaptic has announced a material number of U.S. credit unions as clients over our analysis period, including Guardian Credit Union, Harvard Federal Credit Union, On Tap Credit Union, Patriot Federal Credit Union, People Driven Credit Union, Innovations Financial Credit Union and Wildfire Credit Union.
This area of the U.S credit market exhibits significant momentum and a maturing universe of solution providers. In addition to Scienaptic, we highlight Upstart, Corridor Platforms, VeraScore and Zest AI as vendors of note. Upstart positions itself as a marketplace between borrowers and lenders. Examples of Upstart clients over this period include Idaho Central Credit Union, Kaua’i Federal Credit Union and Naveo Credit Union.
A parallel trend is evident in Asian markets, where underbanked community groups have traditionally been widespread, often with limited access to credit. Singapore-based finbots.ai is a leading provider in the Asian markets with its creditX platform providing credit scorecards using artificial intelligence and behavioural analytics. Examples of finbots.ai client announcements over the analysis period include Baiduri Bank (Brunei), Sathapana Bank (Cambodia), KBZ Bank (Myanmar), Samunnati (India).
Alongside, the deployment of new credit assessment methods, there is momentum in the innovation of new governance, management and control frameworks and tools to ensure credit assessment is carried out in a fair and ethical way. Many of the aforementioned vendors build such governance into their solutions. There is also, however, a sub-sector of the vendor market offering specialist tools to manage this governance and mitigate bias in lending decisions. We make particular note of FairPlay and Stratyfy in this regard. Both vendors are U.S market focused, emphasizing the wider trend for credit model innovation in the U.S. Credit Union and Community Banks sector. Example clients announced by FairPlay for its Fairness-as-a-Service offering over this period include Chime and LendingPoint.
In developing and rolling out these solutions leveraging artificial intelligence within credit risk processes, these vendors have forged an extensive network of partner relationships, typically integrating these credit risk solutions into the wider innovations being developed in the fintech community (thus servicing the end-to-end lending value chain). Over our analysis period, these networks have been significantly extended. Noteworthy examples of partnerships which have been announced by these vendors include: Oscilar, EDOMx, Fintech Galaxy, Worldline, Akoya, MoneyLion, DigiFi, Nymbus, CRIF, Fiserv, Sync1 Systems.
Back
An illustration of how probabilistic AI can be grounded inside deterministic financial operations