Data needs to be cleaned before being able to be fed into ML algorithms. Outliers are deleted, faulty entries eliminated, timestamps aligned, inconsistencies removed, metadata added, and transformed into the proper format or structure. Even while using software tools, data cleaning typically represents up to 80 % of the overall project. If data is available in an OPC UA information model, data cleaning can be virtually skipped, thus cutting up to 80 % out of the duration of the industrial analytics project. Sounds like magic. How is that possible? OPC Unified Architecture (OPC UA) is the de facto standard for data exchange between products from different industrial equipment manufacturers. It allows modelling objects with any degree of complexity. Companies actively using OPC UA in their machines or production line implicitly make data available in a clean, structured way: Not only can they feed their data directly into ML algorithms; they can also put the results back into the OPC UA information model, for use by approved third parties. Hence: Fast ML for Industry by OPC UA in, OPC UA out!