The Close the Loop page (Documentation Index
Fetch the complete documentation index at: https://docs.buildbetter.ai/llms.txt
Use this file to discover all available pages before exploring further.
/success/close-the-loop) — formerly /success/releases — is the account-level view of shipped features against customer requests. Where the Projects close-the-loop card handles one project’s requesters, this page handles every request across every account, matched against your full ship history.
The Four Tabs
Overview
- Summary: matched requests, high-confidence matches, pending draft emails, completed sends
- Analysis progress card — elapsed time, current phase (indexing, matching), processed count
- The job runs once per session by default; a refresh button kicks off a re-run
Requests
A list of every customer feature request in your workspace. Each row shows:- Request title and description
- Status (open / in progress / matched / closed)
- Count of matched releases per request
- Source signal (the call or feedback that captured the request)
Shipped
A catalog of shipped features pulled from your release sources. Organized by release / date. Sources:- GitHub releases & tags — auto-synced from configured repos
- RSS feeds — for products with public changelogs in RSS
- Changelog.com or similar external changelog hosts
- Manual entries — for releases that aren’t tracked elsewhere
Settings
Configure where shipped features come from:- GitHub repos — connect repos to sync releases from
- RSS feeds — add public changelog URLs
- External changelog sources — Changelog.com or similar
- Manual feature entry — add features the auto-sync missed (e.g., “Requested by CEO in call, didn’t make the changelog”)
- Refresh GitHub repos — force a re-sync
The Matching Engine
When you open Close the Loop, an AI job auto-runs to match every open request against every shipped feature. The match uses:- Semantic similarity — embeddings on request text vs. release notes
- Explicit keyword matching — feature names, ticket numbers, slugs
- Confidence scoring — each match gets a 0–1 confidence
Workflow
Open the page; wait for analysis
The analysis progress card shows phases. On a workspace with hundreds of requests, the first run can take a few minutes; later runs are faster.
Review high-confidence matches
On the Overview tab, the matched-requests count gives you the day’s work. Open the Requests tab to drill in.
Draft the email
Click any matched request to draft a templated “we shipped your request” email. The draft includes the request quote (with attribution), the release name, and a link.
Preview and send
Edit the draft, preview, and send via BuildBetter’s email integration. The send is logged for audit.
Manual Entry
Two cases where manual entry matters:- Request the AI missed — type the request into the Requests tab with attribution
- Feature the auto-sync missed — add a shipped feature with a release date and description so the matcher can use it
How It Connects to Projects and Re-engage
- Projects Close the Loop is the per-project view of who asked for this specific project. Success Close the Loop is the cross-project view — every request across every account, against every shipped release.
- Re-engage uses the same matching engine to elevate closed-lost deals where you’ve since shipped the blocker. If a deal lost on “missing SSO” and you shipped SSO this quarter, that account surfaces in Re-engage with an elevated score.
Empty States
- No matched requests yet — usually means the analysis hasn’t run, or you don’t have shipped features connected. Check Settings.
- No shipped features — connect GitHub or RSS, or add manual entries.
- No requests — your workspace doesn’t have customer requests captured yet. They typically come from call signals, feedback, or surveys.