Cody's deep search across large repositories and dual-workflow IDE chat mode is the enterprise developer experience in practice

The global repository context using Sourcegraph's powerful indexing understanding entire codebases rather than just the current session is the capability that makes multi-file questions answerable. Asking "where does the application handle user authentication failures and what are the retry patterns" across a three-million-line codebase and getting an accurate answer with source references is the practical capability demonstration.

The deep search providing results across large repositories that would take a human developer hours to compile manually is the research acceleration that changes the investigation workflow. A five-minute natural language search versus a two-hour manual grep and file review on the same question produces the same information at dramatically different time costs.

The dual-workflow of IDE chat mode for development work and a separate interface for research and documentation being available reflects the different contexts where enterprise developers need AI assistance. Not all questions are code questions.

For senior engineers maintaining large legacy codebases: what type of cross-repository question do you find yourself spending the most manual time investigating?