1. Ease of Use: Google AutoML Tables provides a user-friendly interface that allows users to easily create machine learning models without the need for extensive coding knowledge, while XGBoost requires more technical expertise and knowledge of coding languages like Python.

2. Automatic Feature Engineering: Google AutoML Tables automates feature selection, extraction, and engineering to improve model performance, whereas these tasks need to be performed manually in XGBoost.

3. Scalability and Performance: XGBoost is known for its scalability and performance on large datasets due to its ability to handle a high volume of data efficiently, while Google AutoML Tables may have limitations when dealing with extremely large datasets.

4. Model Interpretability: XGBoost provides greater transparency and interpretability of the model's decision-making process, allowing users to understand how predictions are made, which may be more challenging with Google AutoML Tables.

5. Customization and Fine-tuning: XGBoost offers more extensive customization options and opportunities for hyperparameter tuning to optimize model performance, whereas Google AutoML Tables may have limitations in terms of adjusting specific parameters.

6. Deployment and Integration: XGBoost models can be easily integrated into various production environments, while Google AutoML Tables may have limitations in terms of deployment options and integrations with other systems.

In Summary, Google AutoML Tables and XGBoost differ in terms of ease of use, automatic feature engineering, scalability, interpretability, customization, and deployment options.