The correct option is Provide explicit inline comments with examples and use focused prompts that state the naming rules so Copilot aligns with them.

Copilot generates suggestions from the immediate context and your recent instructions. Inline comments that spell out the naming conventions and show a short example near the cursor strongly bias the model toward your patterns. Focused prompts in chat or in the file that state the required prefixes and casing for methods and services guide the next completions. Reinforcing these patterns across your repositories gives the assistant many signals to follow your style guide.

You can place a brief style note at the top of a file or include a small block of compliant examples and Copilot will mirror those names in its proposals. When you need to tighten the guidance you can ask it in your IDE to produce names that match the rules and include a few correct identifiers for immediate context.

Train a new custom Copilot model for your codebase is not available to customers. Copilot uses shared foundation models and enterprise features can add context but they do not let you retrain or fine tune a private model for your naming rules.

Disable Copilot suggestions in critical folders to avoid noncompliant names does not steer the assistant to follow your standard and only removes help where consistency matters most. Turning suggestions off reduces coverage and leaves the burden on manual enforcement rather than guiding the model.

Rely on IDE code snippets and templates instead of Copilot to maintain consistent names can help with repetition but it does not influence what Copilot suggests and it replaces assistance instead of improving it. Snippets are static and do not adapt to context so they are not the best way to steer the assistant toward your conventions.

Full AWS Practitioner Certification Question