“Data-driven decision-making” is a popular buzz-phrase, but many organizations struggle to encode complex decision processes within software. This is because it is difficult to prove that such systems are taking the correct decisions, unclear how they arrive at them, and risky to alter once they are in production. If the decision process is regulated or safety-critical then the impact of these failures can be drastic.
We develop simple and transparent rules engines designed to separate the decision logic and workflow from the underlying implementation. Our systems allow their users to easily build, verify and picture their decision processes in simple terms, whilst the engine underneath takes care of implementing them.
We developed a decisioning engine to scan millions of customer journeys and determine who may have been mis-sold a financial product
Our client was a major financial institution with a substantial book of financial products that had come under regulatory scrutiny. The client had agreed to conduct a review of past business across several product lines. Any customers deemed to have suffered financial detriment were to be contacted and offered redress. The challenge was that the data volumes were extremely large, the data were of variable quality and the decision criteria were extremely complex.
Our team developed a data processing and decisioning engine that could reconstruct individual customer journeys through the product sales processes. From here it could apply a catalogue of interdependent rules to determine where possible customer detriment had occurred. The system was then able to manage selected customers through the redress journey and produce reports for regulators, proving that all possible cases of detriment had been appropriately treated.
We developed the credit-decisioning engine used by a new entrant to the commercial lending marketplace
Following the 2008 Financial crisis credit lines individuals and small businesses offered by traditional banks dried up. In response a variety of new challengers in the Financial Services space have arisen, offering new forms of credit.
Our client provided invoice factoring services to small businesses. As part of the application process the client had to decide whether the applicant was an appropriate recipient of credit – a decision with both commercial and regulatory components.
The challenge they faced was that such decisioning processes are very complex, and it is difficult to encode such complexity into software in a way that is at once transparent, provably correct and easy to modify. Existing software products (often referred to as rules engines) were found to be either extremely costly or extremely complex to administer.
We developed a credit decisioning engine, run as a real-time service in the cloud. It enabled the client’s analysts to compose complex decisioning logic from collections of simple rules. The system could combine these rules automatically and take decisions on incoming credit applications. Our system had the added advantage of being able explain why any given lending decision had been taken – a capability since enshrined as an applicant’s right by the EU’s new GDPR legislation.