Intelligent WhatsApp Supplier Negotiation Agents: Autonomous RFQ Broadcasting & Purchase Order Generation

1. Why Traditional B2B Procurement Portals Fail with Indian Vendors

Large enterprises frequently attempt to enforce rigid vendor supplier portals, demanding that small and medium raw material suppliers log into a web dashboard, remember complex passwords, and manually type item bids. In the industrial corridors of Rajkot, Morbi, and Ankleshwar, these portals sit empty. Suppliers conduct 95% of their daily trade over WhatsApp, voice calls, and casual chat.

Because procurement managers are bogged down in manual phone calls, comparing hand-scribbled notes, and manually creating RFQs in ERP, procurement cycles stretch across days. Volatile commodity prices (steel coils, virgin plastic granules, kraft paper) fluctuate while quotations sit uncollected.

2. Conversational State Machine Architecture

The conversational procurement agent integrates the official WhatsApp Cloud Business API with an asynchronous Odoo event loop. When raw material levels dip below reorder points, the agent executes the following lifecycle:

1. Target Broadcast: Sends structured WhatsApp messages to certified vendors with item code, quantity, technical specs, and required delivery date.
2. Conversational Negotiation: Handles vendor queries regarding packing requirements, payment credit days, and delivery terms.
3. Structured Extraction: Parses raw messages, PDF quotes, or image snapshots into structured fields: unit_price, freight_terms, payment_terms, tax_inclusive.
4. Bid Normalization: Normalizes prices into net landed cost per unit and populates Odoo purchase.order.line records in status sent.

3. Production Odoo 19 Python ORM Quote Ingestion Blueprint

Below is the Odoo ORM model processing incoming conversational quotations and computing the landed ranking:

# -*- coding: utf-8 -*-
from odoo import models, fields, api, _

class PurchaseOrder(models.Model):
    _inherit = 'purchase.order'

    whatsapp_chat_id = fields.Char(string="WhatsApp Thread ID", readonly=True)
    vendor_raw_quote_text = fields.Text(string="Incoming Vendor Narrative", readonly=True)
    ai_parsed_confidence = fields.Float(string="AI Extraction Confidence", readonly=True)
    landed_cost_estimate = fields.Monetary(string="Calculated Landed Cost", compute="_compute_landed_cost", store=True)

    @api.depends('order_line.price_unit', 'order_line.product_qty')
    def _compute_landed_cost(self, freight_rate=0.04):
        for po in self:
            base_total = sum(l.price_subtotal for l in po.order_line)
            # Add estimated freight and unloading factor
            po.landed_cost_estimate = base_total * (1.0 + freight_rate)

    def process_incoming_whatsapp_quote(self, partner_id, unit_price, lead_time_days, raw_text, confidence):
        """
        Invoked by WhatsApp Webhook Receiver with extracted quotation details.
        Updates or creates draft purchase order lines safely within Odoo ORM.
        """
        self.ensure_one()
        if self.state not in ('draft', 'sent'):
            return False

        self.write({
            'vendor_raw_quote_text': raw_text,
            'ai_parsed_confidence': confidence
        })

        for line in self.order_line:
            line.write({
                'price_unit': unit_price,
                'date_planned': fields.Datetime.now() + fields.Date.timedelta(days=lead_time_days)
            })

        self.message_post(
            body=f"WhatsApp quote received from {self.partner_id.name}: ₹{unit_price:.2f}/unit, Delivery: {lead_time_days} days. Confidence: {confidence*100:.1f}%",
            message_type='comment'
        )
        return True

4. Anti-Collusion Safeguards & Commercial Governance

To ensure absolute integrity, the negotiation agent operates under strict blind-bidding rules: no vendor is ever told another vendor's quoted rate. Furthermore, the agent is strictly prohibited from confirming purchase orders autonomously; it only stages the comparative analysis for the authorized Chief Commercial Officer.

5. Implementation & Business Value Realization

Deploying conversational procurement typically yields a 4-7% reduction in raw material purchasing costs within 90 days simply by expanding the active vendor quoting pool from 2 to 5+ suppliers per procurement requisition.