You are helping me build the following AI agent workflow.

## Goal
Evaluate Vendor Proposal PDFs: Upload a vendor proposal PDF via form; the agent extracts text, summarizes key points, flags risks, logs results to Sheets.

## Specification
- What it does: Automate the evaluation of vendor proposal PDFs by extracting text, summarizing key points, and identifying risks using AI. The system logs all findings in Google Sheets and sends immediate notifications via Gmail for either auto-approval or manual review based on detected risk flags.
- Trigger: Triggered by an event (Manual · on new form submission)
- Autonomy: You approve along the way
- Expected setup effort: about an afternoon
- Tools/services involved:
  - n8n Form: Collect vendor proposal PDFs and submitter details
  - OpenAI Chat Model: Summarize proposals and flag risks via AI
  - Google Sheets: Log evaluation results and risk metadata
  - Gmail: Send approval or review alerts to stakeholders

## Known pitfalls, handle each one explicitly in your implementation
1. Large PDFs may exceed token limits so split the text into chunks before sending them to the model.
2. Generic prompts can miss critical legal clauses so include specific examples of high-risk terms in your system prompt.
3. Gmail API quotas are strict per day so add error handling to retry failed sends or queue them for later processing.
4. PDF extraction quality varies by scanner resolution so verify text accuracy with a sample run before automating fully.

## Reference implementation
https://n8n.io/workflows/16596 (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/evaluate-vendor-proposal-pdfs/ via usecasesforagents.com