AI-assisted features

This page serves as a resource for team members looking for information about working on AI features at GitLab. The first iteration of this page is a list of links, loosely organized. This will improve as we build up our best practices and consolidate/improve our documentation.

Getting started on an AI feature

• Suggest an AI feature: Issue template for suggesting or requesting an AI feature

Design and UX research for AI features

• Pajamas AI guidelines: How to design the user experience for AI features. Includes design patterns and guidelines.
• UX research in AI space: Helps team members evaluate the usefullness and usability of AI features.
• UX maturity guidelines: Suggested UX validation to move up in AI feature maturity.

Evaluating Quality of AI features

• The AI Framework team has guidance for developing, testing, and evaluating your AI-powered feature.
• The UX Research team maintains a usability evaluation and scoring process to ensure ease of use.
• For fast, qualitative feedback on AI response quality, as well as to get an idea of what kinds of prompts users write, you can run a UX Chat Bash.

Special considerations for AI features

• AI ethics principles: Principles to guide our decision making as we continue to build AI features into GitLab and to ensure that these features properly embody our values.
• AI data usage and privacy: Public facing docs about how we use and protect data.

Engineering AI features

• Development process AI features
• List of language models: All the models we are using to power features.
• Process for choosing or switching AI vendors

Support engineering AI features

• Support AI workflow: This page has information for handling tickets related to AI features.

Current AI features and teams

• AI features by tier: All features that are available to use, and how to access them.
• AI teams

View page source - Edit this page - please contribute.