You are helping me build the following AI agent workflow.

## Goal
Customer Feedback Analysis & Response: Captures Jotform feedback, uses AI to analyze sentiment, sends thank-you emails for positive reviews.

## Specification
- What it does: You receive instant thank-you emails from customers who leave positive reviews, while negative feedback automatically alerts your team on Slack so you can address issues quickly. The system captures all Jotform submissions, analyzes sentiment with OpenAI, and logs every response in Google Sheets for easy performance tracking without manual effort.
- Trigger: Triggered by an event (Event · on new Jotform submission)
- Autonomy: Fully hands-off
- Expected setup effort: about an afternoon
- Tools/services involved:
  - Jotform: Collecting customer feedback submissions
  - OpenAI: Classifying sentiment and root cause of feedback
  - Google Sheets: Logging submissions for analytics and tracking
  - Slack: Alerting CX team on negative feedback
  - Send Email: Sending recovery or thank-you emails to customers

## Known pitfalls, handle each one explicitly in your implementation
1. OpenAI API costs can escalate with high volume, so implement caching for identical feedback entries or set a token limit per request to control expenses.
2. Email deliverability may suffer if you send too many messages from a new domain, so warm up your sending IP and include proper SPF/DKIM records in your email configuration.
3. Sentiment models can misinterpret sarcasm or industry-specific jargon, so add a fallback rule that routes ambiguous responses to human review rather than automatic action.
4. Google Sheets has row limits for real-time writing speed, so batch log entries every few minutes instead of writing immediately after each submission to avoid latency spikes.

## Reference implementation
https://n8n.io/workflows/9623 (workflow template)
Fetch and inspect this before building. If it matches my stack, adapt it;
if not, rebuild the pattern with my tools.

## Process requirements
1. Before building: ask me which of the listed tools I actually use and
   what my platform is (n8n / Make / code / other). Do not assume.
2. Adapt the pattern to my answers; do not force the reference stack.
3. Address every pitfall above; tell me how you handled each.
4. Provide a test plan I can run before letting this touch real data.
5. Ask before any step that sends messages, modifies data, or spends money.

Source: https://usecasesforagents.com/use-case/automated-customer-feedback-analysis-response/ via usecasesforagents.com