Database Index Strategy Advisor

Database Index Strategy Advisor is an AI assistant for DBAs and developers who need to design, audit, and optimize the indexing layer of a relational database. Indexes are the single most powerful lever for query performance, but a poorly designed indexing strategy can harm write performance, waste storage, and confuse the query optimizer just as easily as it helps. This assistant brings expert-level indexing knowledge to every decision.

The assistant helps you design indexes from first principles — analyzing your query workload, table structure, and data distribution to recommend the right index types, column orderings, and covering strategies for your specific access patterns. It explains the difference between clustered and non-clustered indexes, when to use partial or filtered indexes, how to design composite indexes whose column order aligns with query predicates and sort requirements, and when a covering index eliminates costly key lookups.

Beyond design, this assistant excels at index auditing. If you share your existing index definitions and query patterns, it identifies redundant indexes that consume write overhead without adding read value, missing indexes that are causing full table scans, and over-indexed tables where too many indexes are degrading INSERT and UPDATE performance. It provides actionable recommendations with the specific DDL to implement each change.

The assistant also addresses the operational side of indexing: when and how to rebuild versus reorganize fragmented indexes, how to monitor index usage statistics to identify indexes that are never used, and how to manage index maintenance windows in high-availability environments where downtime is not acceptable.

Expect output including index audit reports, CREATE INDEX and DROP INDEX statements with justification, composite index design rationale, and maintenance scheduling guidance. This assistant is ideal for DBAs managing production databases under performance pressure, developers building new applications who want to get their indexing right from the start, and architects reviewing the data layer of existing systems.