Salesforce · Rootstock · ChatGPT

Your entire POS flow. One AI conversation.

Talk to us

MCPPOSService connects ChatGPT to your Rootstock ERP via Salesforce Model Context Protocol — letting sales staff take an order from product search to fulfillment without leaving the chat.

10
AI Tools
5
Payment Methods
1
Conversation
0
Manual Screens
Why It Matters

Order management shouldn't require switching screens

Traditional POS flows scatter staff across product catalogues, customer records, order forms, and payment terminals. Every handoff is a place to lose context — and a moment where the customer waits.

Traditional POS flows scatter staff across product catalogues, customer records, order forms, and payment terminals. Every handoff is a place to lose context — and a moment where the customer waits.

MCPPOSService exposes Rootstock's entire order lifecycle as a structured API that ChatGPT can call in sequence. The AI holds the state, asks the right questions, and walks the conversation from product lookup to fulfilled shipment — your staff just talks.

Built on Salesforce's Model Context Protocol with native Apex REST, it slots into any Rootstock org without middleware or third-party integration platforms.

Native Rootstock integration
Reads live inventory, creates soapi staging records, and triggers Rootstock's async order engine — exactly as the native UI does.
Quote-to-order in two API calls
createQuote sets isquote=true; convertToOrder inserts a Change Header and Rootstock's async engine clears the quote flag — matching the “Convert Quote to Order” button behaviour exactly.
Five payment paths, zero guesswork
Cash prepayment, Stripe Terminal tap-to-pay, Stripe Checkout link, direct deposit, and on-account — all token-driven so ChatGPT never constructs payment data manually.
Fulfillment-aware closing
Stock-availability check per order line before fulfillment, with Cash and Carry, Delivery, and Warehouse Pickup methods creating the correct sydata records.
The Complete Pipeline

10 steps. One conversation.

Each tool hands the AI its next instruction. Staff follow the chat; Rootstock follows the API.

Step 01 GET
searchProducts
Keyword search across Rootstock inventory items
Step 02 GET
searchCustomers
Match customer master to retrieve customerId
Step 03 GET
getSitesAndConfig
Fetch proximity-ranked delivery sites & order types
Step 04 POST
createQuote
Staging soapi insert with line items & discounts
Step 05 GET
getQuoteSummary
Itemised calculation with GST tax breakups
Step 06 POST
convertToOrder
Change Header staging promotes quote to Sales Order
Step 07 POST
getPaymentOptions
Generates tokenised payment options for checkout
Step 08 POST
processPayment
Executes tender via Terminal, link, cash, or terms
Step 09 GET
getFulfillmentStatus
Verifies stock availability across all line items
Step 10 POST
fulfillOrder
Creates Rootstock sydata Ship & Invoice records
Tools Reference

Everything the AI can call

Each operation is a registered Salesforce External Service — discoverable, schema-validated, and callable by ChatGPT's tool-use layer.

Step 01
Product Search
GET /MCPos/products
Searches Rootstock inventory by keyword. Returns productId, name, unitPrice, stockOnHand, and a per-site inventory breakdown. The AI surfaces matches and their live stock counts to the customer.
Step 02
Customer Lookup
GET /MCPos/customers
Finds a Rootstock customer master by name or customer number. Returns the Salesforce customerId needed for quote creation. Supports partial-name matching across the rstk__socust__c object.
Step 03
Sites & Config
GET /MCPos/config
Returns delivery sites ranked by proximity to the customer's city and state, plus the available order types. Includes a recommendedSiteId so the AI can pre-select the closest warehouse.
Step 04
Quote Builder
POST /MCPos/create-quote
Creates a Rootstock QUOTE via the soapi staging object (isquote=true). Accepts customerMasterId, orderTypeId, siteId, and a line-items array with optional per-line discount percentages.
Step 05
Quote Summary
GET /MCPos/quote
Retrieves the full quote with all line items, unit prices, quantities, subtotal ex-GST, GST amount, and total inc-GST. The AI presents this as an itemised summary for customer confirmation.
Step 06
Order Conversion
POST /MCPos/convert-to-order
Converts the quote to a live Sales Order by inserting a Change Header staging record (isquote=false). Mirrors the native “Convert Quote to Order” button — Rootstock's async engine clears rstk__sohdr_isquote__c automatically.
Step 07
Payment Options
POST /MCPos/payment-options
Returns tokenised payment options for a live (non-quote) Sales Order. Each option carries a display label and an opaque paymentOptionToken — the AI shows labels to the customer and passes the chosen token to processPayment.
Step 08
Payment Processing
POST /MCPos/process-payment
Executes payment using the token from Step 07. Cash creates a Rootstock prepayment; Credit Card captures via Stripe Terminal; Pay via Link returns a Stripe Checkout URL; Direct Deposit and On Account are recorded without external calls.
Step 09
Fulfillment Check
GET /MCPos/fulfillment-status
Checks available stock for every order line and returns a canFulfillAll flag alongside per-line availability. The AI surfaces any shortfalls before asking how the customer wants to receive the order.
Step 10
Order Fulfillment
POST /MCPos/fulfill
Creates Rootstock sydata Ship and Invoice records for the chosen fulfillment method: Cash and Carry, Delivery, or Warehouse Pickup. This closes the order loop entirely within a single conversation.
Integrated Payments

Five ways to pay, one API call

Payment options are resolved at runtime from your Stripe configuration and Rootstock account settings — the AI never constructs credentials or tokens manually.

Cash
rstk__soppy__c prepayment record
Credit Card
Stripe Terminal — tap or insert
Pay via Link
Stripe Checkout URL, sent to customer
Direct Deposit
Bank transfer, recorded on account
On Account
Net terms, no immediate collection

Bring us the problem you cannot hand off.

We start every engagement with an architectural deep dive, not a generic pilot. Tell us where your enterprise systems stand, and we will map clear, high-ROI AI deployment paths.

Talk to us