Feature Request Triage

Incoming requests from Slack, support, email, and social - deduped, scored, and routed to the right PM area. The next sprint planning works from a triaged tier tree, not a 412-row spreadsheet of 'someone asked for this'.

What the managed workflow does

Pulls from every channel

Slack channels and DMs, support tickets, in-app feedback, forwarded emails, X/LinkedIn mentions - every channel where requests land gets a watcher.

Dedupes fuzzy requests

412 inbound become 187 unique items per week - fuzzy matching on title, body, and sender intent so the same request doesn't get scored five times.

Scores RICE · ICE

Per-request scoring with usage-weight (how many existing customers asked) and persona-weight (which segment) - so the score reflects more than the loudest sender.

Routes by PM area

Growth · Platform · Mobile assignment with the request landing in the right PM's queue, plus a routed reply to the sender confirming where their request went.

Pile → tagger → tier tree · 412/wk inbound to 3-tier triaged

412 REQS/WK · 14 TAGS · 3 TIERS · 3 PM AREAS

Sources requests arrive from

Slack channels + DMsSupport tickets · in-app feedbackEmail · forwarded threadsX / LinkedIn mentions

What lands per cycle

Roadmap-bound · scored to shipBacklog · scored, parkedWontfix · routed to docs / senderPer-area routing · Growth · Platform · Mobile

Sources requests arrive from

Slack channels. Customer-success channels, internal feedback channels, named-account war rooms - every channel where a request might land gets pulled.
Support tickets. Zendesk, Intercom, in-app feedback widgets - request-shaped tickets get auto-triaged to the queue, not buried in support.
Email + forwarded threads. Forwarded customer emails, sales-team field reports, exec inbox - picked up by a dedicated address per area.
X + LinkedIn mentions. Public mentions of the product with a feature ask get logged with sender + segment context.

What lands per cycle

Tier tree. 3-tier outcome - Roadmap-bound · Backlog · Wontfix - with mention counts and PM-area chips per leaf.
Routing replies. Senders get an auto-reply confirming where their request went - closing the loop without burning PM time.
Per-area queue. Each PM sees their queue scoped to their area, ordered by RICE × usage-weight.
Cycle-over-cycle diff. What promoted from Backlog → Roadmap, what de-prioritized - drift visible so the triage is defensible quarter-over-quarter.

What you get, every week