You are helping me build the following AI agent workflow.

## Goal
ECommerce Product Descriptions: Every 15 minutes, it finds pending products in Airtable and uses GPT-4o-mini to generate structured descriptions and features.

## Specification
- What it does: You get fresh, structured product descriptions and features for every pending item in your Airtable base automatically. The system pulls your product attributes, generates the content using GPT-4o-mini, and updates your records so they are ready to go without manual effort.
- Trigger: On a schedule (Scheduled · every 15 minutes)
- Autonomy: Fully hands-off
- Expected setup effort: about an afternoon
- Tools/services involved:
  - Airtable: Fetching and updating product records
  - OpenAI Chat Model: Generating AI content
  - Code: Formatting JSON for Airtable fields
  - Structured Output Parser: Ensuring clean AI output structure

## Known pitfalls, handle each one explicitly in your implementation
1. Batch sizes that are too large can cause token limit errors so keep chunks small enough to fit within context windows comfortably.
2. Invalid JSON structures from the LLM will crash the parser so implement a retry mechanism with strict schema enforcement.

## Reference implementation
https://n8n.io/workflows/11082 (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/automate-ecommerce-product-descriptions/ via usecasesforagents.com