Design an NPS survey for [company type] with the following requirements:

1. Core NPS question (standard wording for benchmarking)

2. Follow-up question asking WHY they gave that score (open-ended)

3. 2-3 specific questions about key experience areas:
   - [Area 1: e.g., product quality]
   - [Area 2: e.g., customer support]
   - [Area 3: e.g., value for price]

4. One question about likelihood to continue using/purchasing

5. Optional demographic question for segmentation

Guidelines:
- Keep total survey under 3 minutes to complete
- Use consistent scale formats
- Avoid leading questions
- Include "prefer not to answer" options where appropriate

Output: Ready-to-implement survey questions with response formats.

TRIGGER: Webhook from e-commerce platform (order fulfilled)

NODES:

1. [Wait] Delay 7 days

2. [Filter] Check if customer hasn't received survey in 90 days

3. [Google Forms] Generate unique survey link with customer ID

4. [Email] Send personalized survey invitation

5. [Wait] Delay 3 days

6. [Filter] Check if survey completed

7. [Email] Send reminder if not completed

8. [Supabase] Log survey send for tracking

TRIGGER: New Google Forms response

NODES:

1. [Google Forms] Get response data

2. [Code] Calculate NPS category (promoter/passive/detractor)

3. [AI Node] Analyze open-ended response:

   PROMPT:

   Analyze this customer feedback response:

   NPS Score: {{ score }}
   Category: {{ category }}
   Open response: {{ response }}

   Extract:
   1. Primary sentiment (positive/negative/mixed)
   2. Key themes mentioned (list up to 3)
   3. Specific product/service mentioned
   4. Urgency level (immediate attention/routine/informational)
   5. Suggested category: product, support, pricing, usability, other
   6. One-sentence summary

   Return as JSON.

4. [Supabase] Store processed response

5. [Filter] If detractor AND urgent → alert flow

6. [Slack] Alert customer success team for urgent issues

TRIGGER: Schedule - Every Monday 6 AM

NODES:

1. [Supabase] Query all responses from past 7 days

2. [Code] Calculate overall NPS and segment breakdown

3. [AI Node] Generate weekly synthesis:

   PROMPT:

   Analyze this week's customer feedback data:

   Total responses: {{ count }}
   NPS: {{ nps }}
   Promoters: {{ promoter_count }} ({{ promoter_pct }}%)
   Passives: {{ passive_count }} ({{ passive_pct }}%)
   Detractors: {{ detractor_count }} ({{ detractor_pct }}%)

   Individual responses:
   {{ response_summaries }}

   Generate a weekly feedback report including:
   1. EXECUTIVE SUMMARY (3 sentences)
   2. NPS TREND (compared to previous weeks if available)
   3. TOP THEMES (most frequently mentioned, with counts)
   4. WINS (what customers praised)
   5. CONCERNS (what customers criticized)
   6. URGENT ITEMS (requiring immediate attention)
   7. RECOMMENDED ACTIONS (specific, actionable)

   Format for executive readability - busy leaders have
   2 minutes, not 20.

4. [Email] Send report to leadership distribution list

5. [Supabase] Archive report for trend analysis

TRIGGER: CRM update - Deal marked "Closed Lost"

NODES:

1. [Wait] Delay 14 days (let emotions settle)

2. [AI Node] Generate personalized outreach:

   PROMPT:

   Write a brief, non-defensive email to a prospect who chose
   [competitor] over us. The goal is learning, not selling. Ask:
   - What ultimately drove their decision?
   - What did the competitor do better?
   - What would have changed their mind?

   Tone: Genuinely curious, appreciative of their time, no hard
   feelings. This is research, not sales.

3. [Email] Send outreach

4. [Supabase] Log for tracking

5. [Wait for reply workflow] Process responses when received

CREATE TABLE feedback_responses (
  id UUID PRIMARY KEY DEFAULT gen_random_uuid(),
  created_at TIMESTAMP DEFAULT now(),

  -- Survey metadata
  survey_type TEXT NOT NULL,  -- 'post_purchase', 'quarterly', etc.
  survey_id TEXT,

  -- Customer info
  customer_id TEXT,
  customer_segment TEXT,  -- For segmented analysis

  -- NPS data
  nps_score INTEGER CHECK (nps_score >= 0 AND nps_score <= 10),
  nps_category TEXT,  -- 'promoter', 'passive', 'detractor'

  -- Open response
  open_response TEXT,

  -- AI analysis
  ai_sentiment TEXT,   -- 'positive', 'negative', 'mixed'
  ai_themes TEXT[],    -- Array of identified themes
  ai_urgency TEXT,     -- 'immediate', 'routine', 'informational'
  ai_category TEXT,    -- 'product', 'support', 'pricing', etc.
  ai_summary TEXT,

  -- Processing
  processed BOOLEAN DEFAULT false,
  actioned BOOLEAN DEFAULT false,
  action_taken TEXT,

  -- Routing
  routed_to TEXT[],    -- Teams this was sent to
  reviewed_by TEXT,
  reviewed_at TIMESTAMP
);

CREATE INDEX idx_feedback_nps ON feedback_responses(nps_category);
CREATE INDEX idx_feedback_created ON feedback_responses(created_at DESC);
CREATE INDEX idx_feedback_themes ON feedback_responses USING GIN(ai_themes);
CREATE INDEX idx_feedback_urgency ON feedback_responses(ai_urgency);