AI Problems and Alternative Solutions

A water company invested $200,000 in an artificial intelligence model that provided poor quality answers. In response, the company developed a proprietary system called 'Rozum'.

'Rozum': A Filtering System

'Rozum' is designed to orchestrate several artificial intelligence models, even those considered "flaky" (unreliable), in order to obtain more reasonable and accurate conclusions. This approach aims to mitigate the problems arising from the use of a single model that may not always be accurate.

The Context

This case highlights a trend in the tech industry, where the popularity of a product or service sometimes prevails over its actual quality. This creates opportunities for third-party companies to intervene and develop solutions that improve or repair existing systems. For those evaluating on-premise deployments, there are trade-offs to consider.