AI Flow Editor
In addition to features for data prepping, performance reporting and data source management, the AI Flow Editor is a crucial part of Trendskout. This editor enables you to design your practical AI applications, ranging from data analysis loading over artificial intelligence and deep learning techniques to output and definition of automation processes. The AI Flow Editor can be accessed through the Trendskout software in your browser, at the push of a button.
This step allows you to connect your data points. First, you can select the type of data input. You can do so using the drag-and-drop feature on the AI Flow Canvas. A variety of input types are supported, including spreadsheets, databases, APIs, Trendskout Pixel and plug-ins. You can combine several data input types: take for example a spreadsheet for the training stage of a classification and API input for the subsequent real-time prediction..
You can connect all input sources from the previous connect stage with the analysis stage. Using a drag-and-drop interface, you can select the type of AI, Machine Learning or Deep Learning you would like to use. A range of different options are available, including labeling-classification, grouping-clustering, text classification, image recognition, prediction and anomaly (or outlier) prediction. Thanks to the built-in selection wizard, you always know what type of analysis to use. There is no need for any coding or behind-the-scenes configuration of the underlying AI and Deep Learning algorithms. Trendskout’s ML will take care of this during the Run & Deploy stage.
Depending on the goal of your application, you can select different automation and output actions. Trendskout allows you, for instance, to send information to a plug-in, send an email or text message, connect to other tools through the Trendskout API, or to generate reports. You can also combine different automation actions simultaneously.
Run & Deploy
Once your AI flow has been configured, you can execute it using the Run feature. This function sets in motion Trendskout AutoML’s Solution Space Exploration, which generates the winning model. After completion, you can move the model to the production stage for live or real-time applications using the Deploy button. You can still make changes to the AI flow for evaluation before you move it to the production stage.