You are helping me build the following AI agent workflow.

## Goal
Multi-Platform AI Sales Agent with RAG: Engages leads across WhatsApp, Instagram, Facebook, Telegram, and web chat using RAG for answers, Airtable for CRM.

## Specification
- What it does: Get 24/7 sales support across WhatsApp, Instagram, Facebook, Telegram, and your website where the agent transcribes voice messages and answers your leads using a knowledge base. It automatically updates your CRM in Airtable and books appointments on Google Calendar so you never miss an opportunity.
- Trigger: Triggered by an event (Event · on new message)
- Autonomy: Fully hands-off
- Expected setup effort: a few focused days
- Tools/services involved:
  - OpenAI Chat Model: Main AI agent reasoning and conversation
  - Postgres PGVector Store: RAG knowledge base for accurate answers
  - Airtable: CRM contact and opportunity management
  - Google Calendar: Appointment booking and scheduling
  - WhatsApp Business Cloud: Omnichannel lead engagement trigger

## Known pitfalls, handle each one explicitly in your implementation
1. WhatsApp template approval can delay deployment so design message structures early and get them approved before launch.
2. Vector search relevance may drop with noisy sales jargon so implement text preprocessing to clean up the knowledge base entries.
3. Context window overflow occurs during long voice transcriptions so truncate history or summarize previous turns before sending to the LLM.

## Reference implementation
https://n8n.io/workflows/4508 (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/multi-platform-ai-sales-agent-with-rag/ via usecasesforagents.com