<\/i>Guest Post<\/strong><\/p>Saverio Lorenzini\u00a0is a Senior Cloud Solution Architect <\/span>with over 25 years of expertise in SQL Server. He specializes in migrations, performance tuning, and\u00a0 optimization. <\/span>This fantastic work that shows a very practical use of AI to help DBAs and DEVs to create better and smarted SQL code, was presented by the author at the Microsoft TechConnect internal conference in Seattle, February 2025, and Saverio was so kind to also agree to publish his work here. Thanks Saverio!<\/div><\/p>\n This article presents an AI-powered approach to automating SQL Server code analysis and refactoring. The system intelligently identifies inefficiencies and common T-SQL anti-patterns, applying best practices through a set of formalized coding rules, using prompt-driven instructions. It not only automatically rewrites problematic and inefficient code but also delivers contextual recommendations to improve quality, security, and maintainability.<\/p>\n Designed to address real-world use cases, this methodology enables organizations to modernize and optimize their SQL workloads more efficiently, accelerating migrations and reducing manual effort in scenarios with large SQL codebases.<\/p>\n The article outlines the architecture of the AI-driven SQL refactoring system and the mechanisms through which it provides intelligent code transformation, with demonstrations highlighting the benefits. The explained technique is a tangible advancement in SQL Server application development and maintenance, reducing manual effort and improving code quality and performance. Sample code is available on GitHub. Link at the end of the article.<\/p>\n The key point of the proposed approach is to identify specific cases of suboptimal SQL code and instruct the AI model on how to refactor them more effectively. The following sections describe key categories of problematic SQL coding patterns that the AI-driven refactoring solution targets.<\/p>\n To illustrate these anti-patterns in practice, the following three examples demonstrate common pitfalls in T-SQL coding and their impact on performance, maintainability, and security.<\/p>\n In the example below, the original query has been rewritten according to two T-SQL development best practices:<\/p>\n![\"Structure](\"https:\/\/devblogs.microsoft.com\/azure-sql\/wp-content\/uploads\/sites\/56\/2025\/06\/Structure.png\")
<\/a><\/p>\nImproving T-SQL code via AI<\/h2>\n
\n
Some Practical Examples<\/h2>\n
Automatic rewrite of SELECT * and ORDER BY n (anti-patterns)<\/h3>\n
\n
SELECT *<\/code>. It retrieves all columns (including potential unnecessary), increasing I\/O and memory. Instead, select only the columns needed later in the code. This reduces data load, improves performance, and keeps code cleaner. In the example, *<\/code> was replaced with only required columns: ProductID,<\/code>\u00a0LineTotal<\/code> and\u00a0 OrderQty.<\/code><\/li>\n