Workflow 1: AI Agent for Restaurant I want to set up an n8n workflow integrating an AI agent capable of automatically handling customer interactions via text messages (WhatsApp) as well as calls (WhatsApp or traditional telephony). Since the restaurant is located in Morocco, the AI agent must be able to communicate fluently and naturally in French and Darija (Moroccan dialect). The agent must have access to a centralized data source containing the restaurant menu (table or database). This source will allow it to: answer questions about the dishes, suggest items, take orders for home delivery. A second table must automatically record orders with all customer information: first name, last name, phone number, address, ordered products, date, and order status. Within the same workflow, the agent must also manage reservations. It should be able to: check availability (time slots and number of seats), propose available time slots, confirm reservations, record the information in a dedicated reservation table or calendar. The objective is to have a centralized system capable of autonomously managing both orders and reservations. Workflow 2: AI Agent for E-commerce (Instagram) I want a second workflow dedicated to handling incoming messages on Instagram. This AI agent must: automatically reply to customer messages, understand voice messages and respond to them in writing, analyze images sent by customers in order to identify the corresponding products in the available catalog. When a customer places an order, the agent must automatically record the information in a Google Sheet including: first name, last name, phone number, Instagram ID, ordered products, date and order status. The agent must be able to detect purchase intent, structure the information, and finalize the collection of the necessary customer details. Possible Improvement: Follow-up and Delivery Tracking System It would be relevant to add an automated post-order follow-up system. This system would be connected to the delivery platform’s API in order to retrieve status updates (in preparation, out for delivery, delivered, delivery failed, etc.). Depending on the status: if the order is delivered, send a confirmation message and possibly request customer feedback, if the delivery fails or remains pending, trigger an automatic follow-up or an internal alert. This would help optimize the successful delivery rate and improve the overall customer experience.