Skip to main content
command CX Codex

synthesize-feedback — Claude Code command: feedback analysis

The synthesize-feedback command turns customer signals into ranked themes and estimates their impact. Use it to prepare evidence and follow-up priorities from feedback.

Apache-2.0 ★ 450 updated 2026-04-03

Install into Claude Code · Manually: put the files in place All methods and platforms ↓
.claude/commands/synthesize-feedback.md
Open on GitHub

What synthesize-feedback does

The command moves feedback through collection, topic clustering, impact modeling, opportunity framing, and packaging for teams. Its outputs include a theme matrix with impact and confidence, a quote bank, an evidence appendix, and action backlog seeds for route-insights.

Set window to choose the analysis period, use personas to filter segments, and select output depth with detail. You can include verbatim quotes and choose ARR, usage, or satisfaction as the impact lens. The command also stores a plan under .claude/plans; data aggregation depends on connected systems being available.

What it helps with

  • group customer feedback into recurring themes
  • estimate how themes affect ARR, usage, or satisfaction
  • collect supporting quotes and evidence for each theme
  • draft priorities and follow-up items for route-insights

Example prompts for the agent

  • Synthesize feedback from the last 60 days for admins and execs in workshop detail: /customer-feedback-orchestration:synthesize-feedback --window 60d --personas "admins,execs" --detail workshop.
  • Group the feedback and prioritize themes by satisfaction impact: /customer-feedback-orchestration:synthesize-feedback --impact-lens satisfaction.
  • Create a concise summary that includes verbatim quotes: /customer-feedback-orchestration:synthesize-feedback --detail summary --include-voice true

How to install synthesize-feedback

Manually: put the files in place

.claude/commands/synthesize-feedback.md

Call it with /synthesize-feedback. For all projects use ~/.claude/commands/

Questions about synthesize-feedback

How do I install synthesize-feedback in Claude Code?

Save the command file at .claude/commands/synthesize-feedback.md.

What does synthesize-feedback do?

It combines quantitative and qualitative feedback, clusters it into themes, and estimates impact. Results include a theme matrix, supporting quotes, and backlog seeds for route-insights.

Which options can I set when running the command?

Use --window for the analysis period, --personas for segment filters, and --detail for output depth. --include-voice controls quotes, while --impact-lens selects the prioritization measure.

Source text

synthesize-feedback.md · Apache-2.0
---
name: synthesize-feedback
description: Consolidates quantitative and qualitative signals into prioritized themes with opportunity sizing.
usage: /customer-feedback-orchestration:synthesize-feedback --window 60d --personas "admins,execs" --detail workshop
---

# Command: synthesize-feedback

## Inputs
- **window** – analysis horizon (30d, 60d, quarter) pulling data from surveys, support, usage, community.
- **personas** – optional filter for personas or segments to highlight.
- **detail** – summary | workshop controls depth of deliverables.
- **include-voice** – true/false to embed verbatim quotes.
- **impact-lens** – arr | adoption | satisfaction to drive prioritization lens.

### GTM Agents Pattern & Plan Checklist
> Mirrors GTM Agents orchestrator blueprint @puerto/plugins/orchestrator/README.md#112-325.

- **Pattern selection**: Feedback synthesis typically runs **pipeline** (aggregation → tagging → impact modeling → opportunity framing → narrative). If tagging + modeling can run in parallel, record a **diamond** block with merge gate in the plan header.
- **Plan schema**: Save `.claude/plans/plan-<timestamp>.json` capturing window, data feeds, dependency graph (CS, product, research, analytics), error handling, and success metrics (theme confidence, coverage %, time-to-synthesize).
- **Tool hooks**: Reference `docs/gtm-essentials.md` stack—Serena for taxonomy updates, Context7 for historic studies, Sequential Thinking for workshop prep, Playwright for dashboard QA when embedding outputs.
- **Guardrails**: Default retry limit = 2 for data pulls/tagging jobs; escalation ladder = Research Lead → CX Leadership → Exec sponsor.
- **Review**: Run `docs/usage-guide.md#orchestration-best-practices-puerto-parity` before workshops to confirm dependencies + deliverables.

## Workflow
1. **Data Aggregation** – collect structured/unstructured feedback from connected systems.
2. **Tagging & Clustering** – apply taxonomy, detect emerging topics, score severity + frequency.
3. **Impact Modeling** – quantify ARR, usage, or sentiment impact with supporting metrics.
4. **Opportunity Framing** – craft problem statements, desired outcomes, and proposed motions.
5. **Narrative Packaging** – produce decks/notes for PM, CS, marketing, and exec consumption.

## Outputs
- Theme matrix with impact, personas, confidence, and owner recommendations.
- Quote bank + evidence appendix tied to each theme.
- Action backlog seeds for `route-insights` command.
- Plan JSON entry stored/updated in `.claude/plans` for audit trail.

## Agent/Skill Invocations
- `cs-analyst` – leads data prep + scoring.
- `research-lead` – ensures insights align with study goals.
- `insight-synthesis` skill – provides framing templates and storytelling patterns.
- `survey-design` skill – highlights methodology caveats.

## GTM Agents Safeguards
- **Fallback agents**: document substitutes (e.g., CS Analyst covering Research Lead) when leads unavailable.
- **Escalation triggers**: if data completeness or confidence metrics miss thresholds twice, escalate via GTM Agents rip-cord and log remediation in plan JSON.
- **Plan maintenance**: update plan JSON/change log when data sources, personas, or impact lenses change to preserve audit history.

---

© gtmagents · Apache-2.0 · gtmagents/gtm-agents

Similar slash commands

All slash commands for Marketing and content →

From the same set224

The whole set →
connect-signals connect-signals sets up data connections, schema alignment, and monitoring for B2B intent, product, and engagement feeds. CX Codex +1 ★ 450repo stars track-source track-source combines partner campaign data and produces attribution reporting with revenue contributions and suggested changes. Use it to review co-marketing results. CX Codex +1 ★ 450repo stars manage-assets manage-assets — a command for orchestrating co-branded asset production, approvals, and distribution across partner teams. Tracks asset inventory, routes review workflows, and packages deliverables for activation. CX Codex +1 ★ 450repo stars optimize-referrals optimize-referrals analyzes referral conversion funnels, fraud patterns, and partner health to generate structured optimization plans in Claude Code. CX Codex +1 ★ 450repo stars define-handoff define-handoff — a command to design lead-to-sales handoff: sets routing rules, response SLAs, and enablement requirements for each sales pod or segment. CX Codex +1 ★ 450repo stars monitor-personalization monitor-personalization audits personalization campaigns, tracks experiment outcomes, and flags consent violations across user segments. CX Codex +1 ★ 450repo stars editorial-ops editorial-ops is a Claude Code skill that plans multi-channel editorial calendars and coordinates GTM content production from ideation through distribution, with KPI guardrails set before assets ship. CX Codex +1 ★ 450repo stars measure-engagement measure-engagement pulls community health metrics, sentiment, and experiment results into one report with a prioritized action list. Use it when you need to see what is happening in a community and what to do about it. CX Codex +1 ★ 450repo stars

Project sponsors

Companies that keep this analytics open