Sage 50 and SAP Ariba integration

What moves between them

The flow runs from SAP Ariba into Sage 50. ml-connector runs Ariba reporting jobs for approved purchase orders and invoice headers, then writes them into Sage 50 as purchase orders and purchase invoices against the matching vendor. Supplier records flow the same direction so new SAP Ariba suppliers become Sage 50 vendor accounts before any document references them, and supplier status can be written back to Ariba through the Supplier Data API where that is enabled. Both systems are pull-only, so the cadence is a scheduled poll, typically aligned to each invoice or purchase cycle, with a high-water-mark timestamp so each run only picks up records changed since the last sync.

How ml-connector handles it

ml-connector stores both credential sets encrypted. On the Ariba side it fetches an OAuth bearer token, caches it for its one-hour life and refreshes before expiry, sends the apiKey header on every call, and appends the realm as a query parameter. On the Sage 50 side it drives a local Windows agent that opens the company file with the Sage user name, password, and Application ID, using a dedicated integration user so it does not collide with the single-session lock when someone has Sage 50 open. Because neither side pushes events, purchase orders and invoices come out of Ariba through async reporting jobs that respect the one-year window per job, and the connector polls on a schedule rather than waiting for a webhook. Suppliers are matched to Sage 50 vendors first, so every purchase order and bill posts against an account that already exists. Ariba rate limits return HTTP 429, so the connector watches the remaining-calls header, backs off, and retries, with automatic retries and a full audit trail on every record so a failed Sage write can be replayed. The agent's SDK version must match the installed Sage 50 year version, and any UK-only VAT fields on Ariba documents are mapped or dropped cleanly.

A real-world example

A mid-sized manufacturing firm of about 150 staff runs procurement and supplier onboarding in SAP Ariba but keeps its books in Sage 50 on a Windows server at the office. Before the integration, an accounts clerk opened each approved purchase order and supplier invoice in Ariba and typed it into Sage 50 line by line, which was slow and produced mismatched vendor names and missed bills at month-end. With Sage 50 and SAP Ariba connected, approved purchase orders and invoice headers are pulled from Ariba on a schedule and written into Sage 50 against the right vendor automatically, and new Ariba suppliers appear as Sage 50 vendor accounts before their first invoice arrives. The clerk reviews exceptions instead of re-keying every document, and the accounts payable ledger stays in step with what procurement approved.

What you can do

• •Pull approved purchase orders from SAP Ariba and create matching purchase orders in Sage 50.
• •Read SAP Ariba invoice headers through async reporting jobs and post them into Sage 50 as bills against the correct vendor.
• •Sync SAP Ariba suppliers into Sage 50 vendor accounts, and update supplier status back in Ariba where enabled.
• •Bridge the Ariba OAuth token and apiKey to the Windows-local Sage 50 SDK login on the on-premise agent.
• •Poll both pull-only systems on a schedule with retries and a full audit trail on every record.

Questions

Which direction does data move between Sage 50 and SAP Ariba?

The main flow is from SAP Ariba into Sage 50. Approved purchase orders, invoice headers, and supplier records move from Ariba into Sage 50 as purchase orders, bills, and vendor accounts. Supplier status can be written back to Ariba through the Supplier Data API where that is enabled, but financial documents are posted only into Sage 50.

Why does Sage 50 need a local agent instead of a cloud connection?

Sage 50 is desktop software with no cloud REST API and no webhooks, so there is nothing to call over the internet. ml-connector runs a small agent on the customer's Windows machine that uses the Sage 50 .NET SDK or Sage Data Objects to read and write records, then syncs to the platform over HTTPS. The machine needs to stay on and logged in for syncs to run.

How does the integration get purchase orders and invoices out of SAP Ariba?

Ariba has no simple endpoint that lists invoices, so ml-connector submits an async reporting job filtered by an updated-date window, polls until the job finishes, and downloads the results in pages. Each job is limited to a one-year window and the platform respects Ariba's rate limits by watching the remaining-calls header and backing off. A saved timestamp means each scheduled run only picks up documents changed since the last sync.