Real-Time SAP B1
Integration

Sales reps had no visibility into customer outstanding invoices, order history, or account status when visiting clients on site.

Sales orders created in the field had to be manually keyed into SAP by back-office staff — introducing delays and the risk of errors.

Invoice PDFs had to be generated separately and shared manually — a time-consuming process that slowed down collections and client communication.

Sales managers had no real-time picture of field activity or pipeline status — decisions were made on stale data days after the fact.

There was no single source of truth between the CRM and the ERP — duplicate records and mismatched data were a constant operational drag.

SAP B1 OData pagination behaves differently from standard REST — $skip and $top interact with server-side session state. We built a paginated fetching engine in Spring Boot using cursor-based pagination with an application-level cache refreshed on a configurable schedule.

SAP B1 uses series-based numbering per document type with prefixes and sequences spread across multiple endpoints. We built a series cache that fetches configuration at startup, maps by document type, and applies the correct prefix logic when generating document numbers.

SAP B1 provides data but not presentation. We implemented a pipeline using OpenHTMLToPDF with Thymeleaf HTML templates — SAP data is fetched, mapped to a document model, rendered to HTML, and converted to PDF. CSS running() patterns handle consistent headers and footers across multi-page documents.

Fetching SAP lookup entities on every request was too slow and created unnecessary load. We built a structured lookup cache layer that preloads at startup, serves from memory at runtime, and exposes a manual refresh endpoint for administrators when master data changes.

SAP's BP object is complex — nested address lines, multiple contact person structures, strict field validation. We reverse-engineered the BP object through the Service Layer metadata endpoint, built a strict field mapping layer with validation before any write, and validated against a SAP sandbox before going live.

SAP Service Layer uses session-based auth with tokens that expire — unlike standard OAuth refresh flows, re-login adds latency if not handled proactively. We built a session manager that tracks session age, proactively refreshes before expiry, and implements automatic retry logic for requests that fail due to timeout.

Manual re-entry between the field CRM and SAP Business One was completely eliminated — data flows automatically, bidirectionally, with no back-office intervention.

Field sales reps gained real-time visibility into customer accounts, outstanding invoices, and order history — directly from their mobile devices, before and during client visits.

Invoices and sales order documents are now generated automatically — branded, print-ready PDFs produced directly from live SAP data, on demand, from the mobile app.

CRM and SAP now share a single source of truth — no duplicate records, no mismatched data, no reconciliation overhead between field and back office.

Sales managers gained a live view of field team activity and order pipeline — no more waiting days for data that was already happening in the field.