Automated machine learning (AutoML) abstracts away most steps in machine learning workflow.
Usually, the system using AutoML will just try different algorithms and choose which of the models gives best predictions. But the system can also utilize meta-learning approaches.
TRADITIONAL WORKFLOW: Data Collection > Exploration > Data Preparation > Feature Engineering > Modeling > Training > Hyperparameter Tuning > Predictions AUTOML WORKFLOW: Data Collection > Predictions
Using AutoML solutions is preferable, if it works for your problem. More complex machine learning problems might require you to get your hands dirty.
Tools and systems that have AutoML features: