Sage 300 and Slack integration

What moves between them

The main flow is pull from Sage 300 into Slack notifications. ml-connector polls Sage 300 on a schedule you define and reads recent transaction batches from Accounts Payable (APInvoiceBatches, APPaymentAndAdjustmentBatches), General Ledger (GLJournalBatches), Accounts Receivable (ARReceiptAndAdjustmentBatches), and Purchase Orders (POPurchaseOrders). These records are formatted as Slack messages and posted to designated channels mapped to their transaction type or approval routing rule. Reference data such as vendor names, customer names, and GL account descriptions flow from Sage 300 master data endpoints (APVendors, ARCustomers, GLAccounts) to enrich the Slack message content so recipients see business context, not just transaction IDs.

How ml-connector handles it

ml-connector stores Sage 300 HTTP Basic Auth credentials (username and password) encrypted and enforces uppercase encoding on every API request per Sage 300 requirements. It stores the Slack OAuth 2.0 bot token encrypted and includes it in the Authorization Bearer header on every Slack API call. On the Sage 300 side, ml-connector polls configured modules on a schedule you choose (e.g. hourly for AP batches, daily for GL journals) using OData filters to retrieve only new or updated records since the last poll, keyed on DocumentDate and time fields. It maps Sage 300 vendor master data to Slack channel names and user mentions so AP invoice alerts route to the accounts payable channel and GL journals route to accounting leadership. Before posting to Slack, ml-connector enforces the rate limit (batching posts if needed to stay under 1 per second per channel) and handles Slack's signature verification requirement by computing the HMAC-SHA256 of the request body using Slack's provided signing secret. Polling records carry a unique transaction ID so duplicate deliveries are idempotent.

A real-world example

A mid-market distributor runs Sage 300 on a company-hosted IIS server for procurement, accounting, and inventory, and uses Slack across all teams for daily communication. Before this integration, the AP team manually checked Sage 300 each morning for new vendor invoices and flagged high-value purchases for approval by email. With Sage 300 and Slack connected, every new AP invoice batch posts to the approval channel within minutes of entry, tagged with vendor name, amount, and due date. The GL team receives a daily summary of journal batches posted the prior day, preventing manual reconciliation work. Purchase order status moves to the procurement channel, so buyers see when goods are received and can trigger invoice matching faster.

What you can do

• •Post Sage 300 AP invoice batches to Slack channels with vendor name, amount, GL account, and due date extracted from the source system.
• •Send GL journal batch summaries daily to accounting leadership channels, including account numbers, descriptions, and batch posting status.
• •Route Sage 300 AR receipt records to collections and credit channels, showing customer name, amount, and invoice reference.
• •Push new purchase order activity (POPurchaseOrders) to procurement channels, including vendor, line-item count, and status.
• •Map Sage 300 master data (vendors, customers, GL accounts) to Slack user mentions and channels so messages reach the right team members and route to the correct channel automatically.

Questions

How does ml-connector authenticate to Sage 300 when credentials must be uppercase?

ml-connector stores the Sage 300 HTTP Basic Auth username and password encrypted and enforces uppercase encoding on every API request as required by Sage 300. The credentials are base64-encoded with the Authorization header, per the HTTP Basic Auth standard. ml-connector validates the credentials against the Sage 300 instance before polling to catch configuration errors early.

What happens if ml-connector sends messages to Slack faster than the rate limit allows?

Slack enforces a 1-per-second-per-channel rate limit for chat.postMessage. ml-connector queues outbound Slack messages and throttles them to stay under the limit, batching multiple transaction summaries into a single Slack message where possible to reduce overhead. If a batch of Sage 300 transactions arrives faster than Slack can ingest, ml-connector queues them with a delay and retries on a backoff schedule.

Does ml-connector support Slack's Events API or only REST polling?

ml-connector uses Slack REST polling methods to retrieve workspace state (users, conversations) and chat.postMessage to send notifications from Sage 300. It does not subscribe to the Slack Events API at this time, so it does not listen for incoming Slack messages or workspace events. All communication is one-way: Sage 300 to Slack.