Intake & triage
Every Friday, every new ask from the form, the request channel and direct asks is matched to a tool, deduped against the live catalog and the Linear backlog, scored, and given an owner, a decision and a reply.
How to read this: the same tool arrives under different spellings ("Brightpath", "brightpath.io", "Bright Path"). Matching folds them into one tool, so the backlog has one issue per tool, not one per message.
Asks per weekly triage, by channel
How to read this: each bar is one Friday triage pass. Because the pass runs every week, no ask waits more than business days for a decision.
Selected tool
Every open tool, ranked
How to read this: click a row to see why it scored that way, the decision, the reply sent and the Linear issue. "Historical backlog" rows are old Linear issues that had no priority or owner until this sweep.
How the priority score works
Five signals, one weighted average
How to read this: each signal is turned into 0–100 points using the buckets shown, then multiplied by its weight. The five results add up to the priority score (0–100).
Does it pick the connectors that got used?
How to read this: we scored connectors launched in the past with today's weights, then checked what really happened to them. Cards: take one connector that got used and one that didn't; how often does the method rank the used one higher? Bars: the share in each tier that reached 50+ organizations using it weekly by day 90. Higher tiers really do get used more.
The rubric
Escalations to the Product Lead
How to read this: partnerships (the vendor runs its own MCP server) and hard calls (strong demand but no API, or a big blocked deal on a low-scoring tool) go to the Product Lead. Everything else is decided in triage.
Linear: the single source of truth
How to read this: the pipeline writes linear_import.csv in Linear's import shape. Existing issues keep their ID and get a priority, owner and status; new tools get new issues. Build briefs for Urgent and High builds go to engineering (example).