Allocation overview
Last updated: March 25, 2026
Understanding how work is categorized across investment categories, and what features are AI-native
AI Investment Mix
AI-native
Span analyzes code diffs and issue metadata to automatically classify each PR and issue into categories like Feature Development, Maintenance, Dev Experience.
Users can override the AI-inferred category assignment.
Labeled Investment Mix
Manual. This version uses explicit Jira tagging or field mapping to categorize work.
Example: issues are labeled as New Feature, Maintenance, Bugs, Tech Debt, etc. in Jira.
Span reads those fields (Epics, Components, Labels, Custom Fields) and rolls up the cost or time investment accordingly.
It’s rule-based, no AI inference.
Workstreams
AI-native
Span analyzes code diffs and issue metadata to automatically create a “word cloud” like topography of initiatives happening within the organization.
This list of workstreams can be optionally customized by mapping Span’s terminology into customer-supplied terminology.
NOTE: This is feature-flagged, please reach out to your Span team to enable workstream mapping.
Cost Capitalization
Manual.
We ingest the list of Epics from Jira that your team is working on.
Teams then tag which epics are capitalizable vs. not. Either in the Span UI, or in Jira (we sync those labels in).
Workstream-centric Cost Capitalization (Beta)
Hybrid. Part AI, part human review.
We use the Workstreams as the foundation for a review process in which managers review time allocation and whether the workstreams are capitalized or not.
This avoids the need for rigorous jira Epic organization. Our AI power is doing the clustering, on top of which capitalization status is set by the user.
Please reach out to your Span team if you're interested in this feature.
FTE-d
FTE-d quantifies effort using FTE days (driven by timestamps of issues/PRs) and optionally converts that to cost if salary-band data is provided.
See here for more information on FTE calculation.