Complimentary access is available for a limited time. Transaction data will be used to improve the system.
WHAT WE OFFER
Outperforms leading competitors where it matters most: debt, gambling and income.
Trained on gold-standard human annotation of over 7 million proprietary transactions.
Seamlessly upload .csv data via our encrypted pipeline, with customisable data deletion.
*Based on in-house testing across 300,000+ transactions.
NOGGIN CATEGORISATION
Our team will be happy to schedule a bespoke session. For a limited time, partners are also invited to join our beta test and receive complimentary transaction categorisation.
LEARN
Our categorised training data of 4,000,000+ transactions was painstakingly annotated by expert annotators, to create a human-verified gold-standard for our automated system to match. The system uses “large-language models (LLMs),” the technology underlying generative AI, to extract information not only from important words, but words and phrases in context, and their interaction with transaction amounts.
Yes. All transaction data is encrypted and stored on secure servers. Only necessary members of the data science team will have access to transactions, and we handle all data in accordance with UK and EU GDPR regulations. We can also accomodate bespoke requirements on request.
No. The categorisation engine assigns all categories as accurately as possible. However, the forthcoming Noggin proprietary credit scoring model will identify the most important categories for predicting the creditworthiness of borrowers.