Case study · Manufacturing and distribution
Automating the order-to-invoice workflow for BBCP across Acumatica, SAP Ariba and SharePoint
BBCP has automated a connected order-to-invoice journey across MYOB Acumatica, SAP Ariba and SharePoint. AI-assisted extraction turns incoming purchase orders into matched sales orders, warehouse documents are captured automatically, and invoices are submitted back into SAP Ariba — cutting processing time by around 90% and saving roughly $150,000 a year.
15–20 min
Manual processing time per purchase order before automation
Baseline reported by BBCP prior to automation.
>90%
Reported reduction in order-processing time
Reduction reported by BBCP following implementation.
~$150,000
Estimated annual savings in processing time
Reported by BBCP following implementation.
7-step
Automated journey from PO receipt to invoice
PO received, AI extraction, sales order creation, Ariba order confirmation, warehouse docs scanned, shipment updated, invoice created.
The challenge
A manual, system-hopping order-to-invoice process.
BBCP's order-to-invoice process depended on manual handling between Outlook, MYOB Acumatica and SAP Ariba, with hours of manual matching every week.
Before automation, an admin would manually download each PO attachment, read the PDF, look up the customer and part numbers, determine the correct branch and warehouse, then create and attach the sales order in Acumatica by hand — around 15–20 minutes per order.
This manual, human-dependent process was error-prone and did not scale: warehouse teams received orders only after the full manual chain was complete, delaying fulfilment.
There was no consistent audit trail across the disconnected systems, and invoice submission back into SAP Ariba was a separate manual step.
- System of record
- MYOB Acumatica
- Services applied
- Workflow automation · Practical AI · System integration
Before and after
The same order, two very different journeys.
What a purchase order used to go through by hand, and the automated path it takes now.
Before · Manual process
Purchase order email received with attachment
Admin downloads attachment
Reads PDF manually
Looks up customer
Looks up part numbers
Determines branch & warehouse
Creates sales order in Acumatica
Uploads attachment to sales order
Warehouse receives order (delayed)
15–20 minutes per PO · Human dependent · Manual errors · Limited scalability
After · Inturing automated workflow
Email arrives
orders@bbcp.com.au
Power Automate trigger
Separate attachments, convert to PDF
AI extracts PO details
- Customer
- PO number
- Line items
- Quantities
- Delivery address
Business rules & matching
API calls to Acumatica
Customer matchPart matchBranch & warehouse matchException / no match? Requires review
YESTeams approval (human in the loop)NOStraight throughCreate sales order via Acumatica API
Attach original PO to sales order
Sales order available in Acumatica
Warehouse receives order immediately
The approach
AI-assisted extraction, validated before it reaches the ERP.
The design principle was simple: automation handles the standard path completely, and people only see the records that genuinely need a decision.
- 01
AI-assisted purchase order extraction
Incoming PO emails to BBCP's orders inbox trigger Power Automate Desktop, which separates attachments, converts them to PDF, and uses AI Builder to extract structured order data — customer, PO number, line items, quantities and delivery address — from unstructured documents.
- 02
Three-way matching against Acumatica master data
Extracted details are matched against Acumatica via API calls: customer matching (handling real-world variants, such as recognising 'Rio Tinto Pty Ltd' as 'Rio Tinto Resources'), part/vendor matching, and branch and warehouse determination — all before a sales order is created.
- 03
Exception handling stays human
Where no confident match is found, the order is routed to a Teams approval flow for a person to review and resolve. Clean, matched orders flow straight through to sales order creation, with the original PO automatically attached.
- 04
Connected through to fulfilment and invoicing
Once confirmed, the sales order is created via the Acumatica API and made available to the warehouse immediately. Warehouse pick lists and kit assembly documents are scanned to SharePoint and linked back to the job, with shipment details updated automatically in Acumatica.
- 05
Invoice submitted back into SAP Ariba
At the end of the journey, the invoice is created and released in Acumatica, and Power Automate Desktop submits it into SAP Ariba — closing the loop from purchase order to invoice without manual re-entry.
Results
Measured outcomes, and general benefits.
We keep these separate deliberately. The first list is what BBCP reported. The second describes what this approach typically delivers and is not a client measurement.
Measured — reported by BBCP
- Processing time per purchase order was cut by around 90%.
- BBCP estimates savings of roughly $150,000 a year in processing time.
- A full audit trail now exists across every transaction, from PO receipt through to invoice submission in SAP Ariba.
- Warehouse teams receive confirmed orders immediately, rather than waiting on a manual chain of lookups and data entry.
General benefits — not client-verified measurements
- Human review is retained exactly where it adds value: sales order review and release, warehouse physical processing, and invoice review and release.
- Order entry capacity no longer scales purely with headcount.
- Staff attention is directed to exceptions rather than routine data entry and matching.
- The process is documented and supportable rather than dependent on one experienced operator.
Is order processing your bottleneck too?
If your team re-types orders into an ERP, a 30-minute discovery call is usually enough to size the opportunity.