Workstreams: AI-Powered Allocation

Last updated: October 27, 2026

Workstreams are AI-found groups of related engineering work, such as "Mobile checkout" or "Data platform migration." Span builds them from your epics, issues, and pull requests, then shows how much effort each one received.

Why it matters

Workstreams show where engineering effort actually goes, even when your Jira or Linear structure is incomplete. You don't need to tag anything first.

  • For engineering leaders: see which initiatives are taking capacity, which are growing, and where only one or two people carry a critical area.

  • For finance and planning: map workstreams to your own project list and reuse that effort in cost capitalization.

How it works

Span groups work with AI once a week. Span looks at the last 90 days of epics, issues, pull requests, and projects, groups related items, and names each group. It writes a short description for each workstream, marked "Generated by AI."

Workstreams have two levels.

  • Top-level workstreams are major areas of effort, such as Mobile App Development, Infrastructure & DevOps, or Data Platform.

  • Sub-workstreams break a top-level workstream down, such as iOS Development and Android Development under Mobile App Development.

  • Unlinked holds effort Span couldn't place in a workstream, most often pull requests with no linked issue.

Span measures effort in FTE-days. Span infers effort from activity: commits, merged pull requests, reviews, and issues In Progress. Nobody logs hours. Each person's active day is worth 1.0 FTE-day, split across the work they touched, and each share goes to one workstream. See Understanding FTE Days in the Workstreams Report for the full calculation.

What the report shows

  • Activity grid: each workstream's name and description, FTE-days, weekly activity, % of total, the teams contributing to it, and cost when cost estimates are on.

  • Activity growth treemap: box size is FTE-days and color is the growth band: Growing fast (above +60%), Growing (+25% to +60%), Stable (−25% to +25%), Shrinking (−60% to −25%), or Shrinking fast (below −60%).

  • Workstream detail page: the AI summary, effort over time, sub-workstreams, contributors by person and by team, and the issues and pull requests in the workstream.

Setup

Workstreams need a connected Git host and issue tracker. Admins can turn the report on or off under Settings → Investment → Choose Your Investment Views.

View your workstreams

  1. Open Insights → Allocation → Workstreams.

  2. Choose Month, Quarter, or the last three months (the default).

  3. Switch between Raw workstreams (Span's groups) and Projects (your mapped project list).

  4. Use the work filter to show the Top 80% of Work (the default), another cutoff, or Show All Work.

  5. Open Display settings to turn on Show Cost Estimates, or to pick the Labeled investment area shown on each workstream's page.

  6. Click Generate Report to download a CSV of workstream data for all teams for a month or quarter. This isn't available in the Projects view.

Rename or merge a workstream

  1. Open the menu on a workstream's row or detail page.

  2. Choose Edit to change the name or description, or Merge into to fold it into another workstream at the same level.

  3. Confirm. Span records renames and merges in your organization's audit log.

Map workstreams to your projects

Admins with the Manage Investment permission see a Refine button. Refining lets you create refinement periods, group people into cohorts, define your project list, and map workstreams to projects. See Allocation Guide: Refining Your Workstreams.

Things to know

  • Workstreams update weekly. The report shows data up to the latest weekly run. A banner tells you when the selected month or quarter isn't complete yet.

  • New workstreams can shift. A workstream younger than about 30 days can change between weekly runs until someone edits or maps it.

  • You can't create workstreams by hand. You can rename, re-describe, and merge them. For your own categories, map workstreams to projects or use Labeled investment.

  • Merges are permanent. Merging moves the workstream's activity into the target and can't be reversed. The Unlinked row and workstreams that come from your mapped projects can't be edited or merged.

  • Effort is days, not hours. A person's active day counts as 1.0 FTE-day no matter how long they worked. Days with no activity count zero.

  • Cost needs salary settings. Cost columns appear only after an admin enters salaries under Settings → Investment → Cost Estimates, and only for people allowed to see cost estimates.

Worked example

A workstream shows 50 FTE-days over a 10-week (70-day) range. That's 50 / 70 = 0.71 FTE: on average, the equivalent of about 0.7 of one full-time engineer worked on it every day of the range. If its weekly average in the second five weeks is 6.0 FTE-days against 4.0 in the first five, it grew 50% and lands in the Growing band.

Tips

  • Rename early. AI names are a starting point. Use your organization's terms so people recognize them.

  • Merge duplicates. Combine workstreams that describe the same effort to keep the list short.

  • Check contributor mix. A critical workstream with one or two contributors is a knowledge-concentration risk.

  • Watch growth bands. Shrinking effort on a priority initiative is worth a question.

  • Reduce Unlinked work. Linking pull requests to issues moves more effort into named workstreams.

Related articles

Understanding FTE Days in the Workstreams Report

Allocation Guide: Refining Your Workstreams

Allocation overview

Introduction to Cost Capitalization

Need help?

Contact your Span account team or reach out to support@span.app.