A business wants to understand the risk and trust levels of AI and start preparing their data to be AI ready. With Solidatus they gain more trust in the source and use of datasets, full visibility into where AI is being used and which systems use data from AI. Additionally, as their teams tag the AI models and datasets with risk and trust scores, Solidatus does calculations that help them understand their trust levels and risk.
A different company has a live AI model that has been deployed to help calculate mortgage products. When someone makes a change, the AI model begins making poor decisions. The company discovers that the AI model had been trained on a dataset with the address input form with 3 lines. A department changed it to 2 lines, unaware of the impact, so the model didn’t know how to interpret it. How Solidatus helps:
- By assigning ownership, they can trace back to the relevant system and resolve the issues with the owners
- It helps document the complete flow, allowing teams to trace back from the AI model use case to understand its lineage
- By using the bi-temporal view, which can show the future impact for all such decisions, the business would have been able to see the impact on AI models prior to making the change – thus preventing the failure.
