How does Span define allocation categories like new features and maintenance work?

Span offers two approaches to classify your development work into different categories:

AI Investment Mix (Automatic, AI-Powered)

Span's AI automatically analyzes your code diffs and issue metadata to classify work into three categories:

• Feature Development - New functionality and user-facing value creation

• Maintenance - Bug fixes, performance improvements, stability, security updates, infrastructure work, operations, and support

• Developer Experience - Refactoring, code style, CI/CD enhancements, monitoring, tooling, documentation, and tests

The machine learning model examines code changes, PR/issue content, commit messages, and historical patterns to make these determinations automatically. You can override the AI's classification if needed.

Classic/Inferred Investment (Manual, Rule-Based)

This approach uses explicit Jira tagging and rule-based matching. You define rules that evaluate issue attributes such as:

• Issue type (Bug, Story, Task)

• Labels and custom fields

• Epic/Initiative hierarchy

• Priority and other Jira metadata

This method gives you more direct control over how work is categorized by setting up specific rules based on your team's existing Jira workflow and tagging conventions.