The problem
Our client’s credit risk underwriting team applied credit lines to counterparties with delayed payment options. Credit amounts and payment terms were heavily influenced by each counterparty’s financial condition, which was analyzed quarterly by the underwriting team. Assessing a single counterparty consumed multiple hours, and given the client’s scale, this presented significant savings potential.
Industry: Oil & Gas
Technology: AI Agents, Neural networks, Python, Cloud
Our solution
We created two integrated solutions that worked together to assess counterparty financial conditions comprehensively.
Qualitative model
Our solution scanned multiple APIs and external databases to identify all news related to the analyzed companies. A custom model then assessed news sentiment in the context of our client’s specific needs and extracted key information that contributed to an overall sentiment ranking for each company during the given period.
Quantitaive model
We developed an automated program to gather quantitative information from external APIs and websites, including company revenue, debt levels, employment dynamics, and other financial metrics. This data fed into our machine learning algorithm, which classified whether each company’s financial condition had improved, declined, or remained stable compared to the previous period.
The results from both solutions were synthesized into a single comprehensive report for each company, complete with data sources and final recommendations for credit line adjustments.
The impact
The client achieved six-figure annual savings by automating a significant portion of the underwriting workflow. Additionally, the system enabled on-demand assessments, increasing the client’s flexibility in scheduling counterparty evaluations.
