SAP Business One and Zendesk integration

What moves between them

The primary flow is SAP Business One into Zendesk. BusinessPartners are synced as Zendesk organizations and users; customer contacts and vendor contacts create the main user population, mapped to their parent organization (customer or vendor) in Zendesk. Email addresses, phone numbers, and postal addresses flow from BusinessPartner master data. On a regular schedule, ml-connector polls SAP Business One for new and updated BusinessPartners and pushes them to Zendesk to keep contact details and company status current. The reverse direction (Zendesk into SAP Business One) is not supported, since Zendesk has no finance data to send back.

How ml-connector handles it

ml-connector stores SAP Business One credentials encrypted and manages session token reuse to avoid the 5-second cold-start overhead of authentication on each request. When a session token expires (HTTP error -5002), a new one is obtained and the failed request is retried. Zendesk OAuth credentials are also encrypted; access tokens are cached until revocation and refreshed as needed. On the SAP Business One side, ml-connector polls the BusinessPartners endpoint with an OData $filter on UpdateDate to fetch only recently changed records, respecting the default page size of 20 and using $skip to iterate through larger result sets. SAP Business One does not publish idempotency key standards, so ml-connector checks for duplicate DocNum values before inserting to prevent re-syncing the same partner record. On the Zendesk side, ml-connector maps each BusinessPartner to an organization and creates or updates associated user records for all linked contacts. If SAP Business One webhooks are enabled and accessible, ml-connector can optionally subscribe to Create and Update events to reduce latency; otherwise polling remains the default. Retries use exponential backoff on network failures, and every record carries a full audit trail with timestamps and error details.

A real-world example

A growing B2B software company runs SAP Business One for order entry, invoicing, and AR. Support requests come in through Zendesk. Before integration, when a customer called in with an issue, support agents had no visibility into whether the account was current on payments, what they had ordered, or what their invoice status was; agents had to switch to SAP Business One manually or ask the finance team for context. With SAP Business One and Zendesk connected, each support ticket automatically shows the customer's organization record with current invoice total, largest order, and account status pulled live from SAP Business One. Support can now resolve issues faster because the ERP context is already in the ticket.

What you can do

• •Sync BusinessPartners from SAP Business One into Zendesk organizations, keeping customer and vendor master data current.
• •Map contact persons from SAP Business One to Zendesk users, so every relevant person at a customer account is available in support conversations.
• •Handle SAP Business One session token expiry and reuse tokens to minimize authentication latency on polling.
• •Encrypt OAuth and session credentials, retry failed syncs with exponential backoff, and maintain a complete audit trail of all records moved.
• •Poll SAP Business One on a schedule you define, with support for webhooks where enabled, to keep customer and vendor details aligned between the two systems.

Questions

Which direction does data move between SAP Business One and Zendesk?

The primary flow is SAP Business One into Zendesk. BusinessPartners (customers and vendors) and their contact persons are polled from SAP Business One and synced as organizations and users in Zendesk to enrich support tickets with ERP context. Zendesk support activity does not flow back into SAP Business One, since Zendesk has no finance or operational data that SAP Business One needs to consume.

How does ml-connector handle SAP Business One session tokens and avoid authentication overhead on every request?

ml-connector obtains a session token via POST /Login and reuses it for multiple API requests, avoiding the 5-second cold-start cost of authentication on each call. When a token expires (signaled by HTTP error -5002), ml-connector requests a fresh token and automatically retries the failed request. All credentials are stored encrypted in the database.

Does the integration work with SAP Business One webhooks or does it rely on polling?

Polling is the reliable default; ml-connector polls SAP Business One BusinessPartners with an OData $filter on UpdateDate to fetch only changed records on your schedule. If you have enabled the Webhook Messenger Service in SAP Business One v10.0 FP 2602 or later, ml-connector can optionally subscribe to Create and Update events to reduce latency, but polling remains the fallback for reliability.