Product Surfaces

Analytics

Read the current Analytics preview without treating board-derived signals, comparisons, or recommendations as final operational truth.

Last reviewed

Availability and purpose

Analytics is Coming Soon. The shell preview is designed to show how Pulse AI can turn workspace activity into an operational summary. It is not yet a public reporting system, and its comparisons, recommendations, and export action should not be used as an audited business record.

The preview is labeled Engineering Intelligence inside the extension. Its purpose is to help an operator notice flow, load, review pressure, and delivery ownership. The public product registry keeps the surface gated until its signals and recommendations are verified as a complete launch workflow.

Use the preview to understand the intended information architecture. Use the board, review records, and completed deliverables as the current source for decisions.

What the dashboard represents

The dashboard groups activity into several views:

  • Review Pressure summarizes work waiting for or moving through review.
  • Workload presents the priority mix of visible action items.
  • Recent Output summarizes recent workspace activity.
  • Activity Intensity shows created and completed items across a recent time window.
  • Operating Mix separates work by lifecycle status.
  • Pulse Score gives a compact completion-oriented reading.

Analytics also has an explicit empty state. When the workspace does not yet have action items, reviews, or Workday output, the page says that there is no analytics data yet and asks the operator to create activity first. That is preferable to invented sample results.

How Pulse Score is calculated

The preview describes Pulse Score as the percentage of board items marked done:

done items ÷ total items × 100

A score of 50 or higher is labeled stable in the current interface; a lower score is labeled warning. That threshold is a presentation rule, not a universal measure of workspace health. A high percentage can still hide overdue work, weak evidence, or a large backlog that was never captured. A low percentage can be normal at the beginning of a substantial project.

Read the score with its status breakdown: done, active, blocked, and review. Review remains distinct from done, so work waiting for founder approval must not inflate the completed count.

Interpreting the signals safely

Start with the underlying question instead of the chart. If review pressure rises, inspect the Review Queue and the affected action items. If workload appears concentrated in high priority, check whether those priorities share a deadline or are stale. If recent output drops, confirm whether the Workday was intentionally paused before treating it as a problem.

Do not compare workspaces as if their board practices are identical. A workspace that captures every small action will produce different totals from one that tracks only major outcomes. Date-range controls change what appears in the dashboard, so record the visible range when sharing a screenshot.

The Export Report control is part of the preview design. Do not treat it as a verified reporting handoff until the public launch guide documents the export format, source range, privacy behavior, and recovery path.

When the numbers look wrong

If Analytics is empty, confirm that the workspace has action items, reviews, or Workday output. Then reopen the page so it can request current data.

If the score disagrees with the board, inspect the counts by lifecycle state and confirm that review items are not being treated as done. If activity appears stale, reload the extension host and compare the visible date range with the timestamps on the underlying items.

If a recommendation appears without enough evidence, do not act on the recommendation alone. Open the source action items and review records, capture the mismatch, and use the support path with a redacted screenshot and the affected date range.

Related links