We Had 300 Jira Epics. Nobody Knew Which Ones Actually Mattered.

Without business value scoring, every epic looks equally important. FlowSquad scores each epic 0–100 and surfaces what moves the needle — cutting PI planning from 2 days to half a day.

What Is Backlog Debt and Why Does It Kill Velocity?

Most product teams are familiar with technical debt — code that was written fast and needs to be cleaned up later. Fewer talk explicitly about backlog debt: the accumulation of epics, features, and stories that were created but never properly refined, scored, or prioritised.

Backlog debt manifests in predictable ways:

• 3-year-old epics that were valid ideas in 2021 but are now irrelevant — yet still appear in planning discussions because nobody officially closed them
• Duplicate epics from different product managers who weren't aware of each other's work (common in teams larger than 15)
• Context-free epics — "Improve search performance" with no description, no acceptance criteria, and no record of why it was created
• Inconsistently sized epics — some are "build entire payment system" sized, others are "fix button colour" sized, making relative prioritisation meaningless

The cost of backlog debt isn't primarily in the backlog itself — it's in the planning overhead it creates. Every sprint planning session that includes a discussion of "which of these 300 epics should we work on this sprint?" is a session where your most senior, most expensive people are doing archaeology instead of planning. Research by McKinsey puts the cost of poor prioritisation at 20–40% of engineering capacity wasted on low-value work.

How Do You Extract Epics from Jira + Confluence Automatically?

FlowSquad's Jira integration connects directly to your Jira instance (Cloud or Server) and imports your full epic list with metadata: title, description, creation date, assignee, label, component, and linked stories. The import is read-only by default — FlowSquad never modifies your Jira data without an explicit push action.

The Confluence integration complements Jira by pulling context from linked Confluence pages — PRDs, business cases, OKR documents, and strategy pages that provide the background needed for meaningful scoring. FlowSquad automatically cross-references Jira epic descriptions against linked Confluence content to build a richer understanding of each epic's intent and business context.

# FlowSquad Jira integration configuration
jira:
  server_url: "https://yourcompany.atlassian.net"
  auth_type: "api_token"        # or "oauth2"
  api_token: "${JIRA_API_TOKEN}"
  project_keys:
    - "ENG"
    - "PROD"
    - "PLATFORM"
  import_types:
    - epic
    - story
    - feature                   # SAFe feature type
  import_filters:
    status: ["Open", "In Progress", "Backlog"]
    max_age_days: 730            # ignore epics older than 2 years
  confluence:
    base_url: "https://yourcompany.atlassian.net/wiki"
    auto_link_pages: true        # auto-discover linked Confluence pages

How Does FlowSquad's Business Value Scoring Work?

FlowSquad's business value scoring engine evaluates each epic against five dimensions, producing a composite 0–100 score:

The scoring is not a black box — FlowSquad provides a per-dimension breakdown for every epic, showing exactly why it received its score. Product managers can adjust dimension weights to reflect their organisation's current priorities (e.g., increase regulatory compliance weight during an audit period).

What Does SAFe Program Increment Planning Look Like with Scored Epics?

SAFe Program Increment (PI) planning is the biannual event where multiple Agile Release Trains (ARTs) align on what they'll deliver in the next 10–12 weeks. In the typical pre-FlowSquad world, PI planning involves days of preparation: manually assembling feature lists, debating priorities in a room of 50+ people, and producing a plan that's already out of date by the time it's published.

With FlowSquad's scored epic backlog, PI planning becomes dramatically more structured:

1. Pre-PI: Product managers run FlowSquad's batch scoring across all open epics, producing a ranked list with scores and rationale. This takes 45–90 minutes instead of 2–3 days of manual preparation.
2. PI Event: The ranked list is the starting point — not a blank slate. Discussions focus on the top 20% of epics (those scoring 70+), and the bottom 40% are pre-filtered out, reducing the decision surface by 60%.
3. Post-PI: Accepted epics are pushed back to Jira as Programme Increment-tagged epics with their FlowSquad-generated stories pre-attached. No re-entry required.

One enterprise team using FlowSquad for SAFe PI planning reported reducing their PI preparation time from 3 days to 6 hours and their PI event itself from 2 days to 1 day — saving approximately 400 person-hours across a 50-person planning event.

From 300 Epics to a Prioritised Roadmap in One Session

The practical workflow for a product manager inheriting a 300-epic backlog:

1. Connect Jira and Confluence in FlowSquad's 10-step setup wizard (10–15 minutes)
2. Run a batch import — all 300 epics pulled into FlowSquad with their descriptions and linked pages
3. Run batch scoring — FlowSquad scores all 300 epics in parallel (~8 minutes for 300 epics)
4. Review the ranked output: typically ~30 epics score 70+, ~80 score 40–69, ~190 score below 40
5. Archive the bottom tier (below 20), defer the middle tier, commit the top tier to roadmap
6. Export the prioritised roadmap to Confluence or push the top epics back to Jira with priority ranks set

The total session takes 2–4 hours for a product manager working alone — including time to review and adjust scores where domain knowledge overrides the AI recommendation. The result is a defensible, documented prioritisation with rationale for every decision.