BT Group deployed a new machine learning platform in its own environment that reportedly reduces the time required to produce an AI product from six months to only six days.
The platform, AI Accelerator, is in charge of continuously checking on the wellness and effectiveness of the artificial Intelligence models created by BT’s creative team. According to BT, AI Accelerator fast-tracks the technical and administrative steps required to ensure a use case or AI product is ready for implementation.
BT AI Accelerator
The platform provisions standardized data processing and tools. Its functionality accelerates the adoption of secure AI models throughout the organization. Due to streamlined software updates and monitoring, the company’s data analysts and scientists free up large amounts of time, according to BT.
In terms of performance, the new platform highlights any ‘drifts’ from baseline norms when the model evaluates data. For instance, this new platform can trigger alerts when analytical performance or accuracy decreases.
One use case, the report notes, is the development of models that calculate the chance that a specific customer will stop using a company’s service within a particular time frame.
The use of AI is sometimes questioned for potential biases and unethical practices. However, BT ensures that all models are made most ethically and safely, with privacy as a top concern. Parameters will be incorporated into the tool’s monitoring feature to ensure that the AI models adhere to the business’s technology standards.
The AI-driven technology used in BT’s products is supported by the AI models created by the company’s development teams, which draw data from a 29-petabyte estate referred to as its ‘digital brain.’
AI Accelerator allows BT to constantly monitor the AI models used through its environment, highlighting any deviation from the baseline norms that determine how AI should generate insights from data.
As a result, instead of maintaining each AI model manually, analysts free up time to concentrate on upcoming data science initiatives, encouraging innovation and speeding up the usage of AI throughout the company.