In the context of digital advertising, accurate data is crucial for optimizing AI models and performing effective historical data analysis. Consider you have a dataset comprising ads data, and you need this data for two primary purposes: to serve AI models and to analyze historical trends. A significant aspect of data preparation is identifying longtail and outlier data points, which can potentially skew the analysis and the performance of AI models. To ensure the highest quality of data, you aim to cleanse the data in near-real time before integrating it into your AI models. What actions should you take to achieve this?