Process a machine-readable PDF invoice with Sage Document Intake
Detailed client guidance covering process a machine-readable PDF invoice with Sage Document Intake, including prerequisites, exact controls, verification evidence, failure boundaries and safe recovery steps for the shipped Sage connector.
Where to goSage Accounting Automation → Document Intake
Before you startConfirm the intended tenant, remote account or invoice context, and your Document intake access. Keep credentials and tokens out of screenshots, tickets and exported evidence. If a previous write may have reached an external service, verify remote state before retrying.
What you’ll accomplish
Detailed client guidance covering process a machine-readable PDF invoice with Sage Document Intake, including prerequisites, exact controls, verification evidence, failure boundaries and safe recovery steps for the shipped Sage connector. This article is intentionally scoped to the behaviour enforced or exposed by the supplied module and does not treat provider marketing features as implemented integration capabilities.
How the workflow fits together
The connector keeps native CRM records authoritative while mapping verified Contact, Sales Invoice and Contact Payment evidence in Sage. Queue, exception, API, reconciliation and compliance records are separate evidence layers, so each material action should be verified at both the native and remote sides.
Follow these steps
Open Sage Accounting Automation → Document Intake.
Confirm you are in the intended tenant and connected business and that your role has the required Document intake access.
Confirm Document Intake is enabled and the PDF comes from an authorized business source.
Upload within the configured file-count and file-size limits.
If the PDF is scanned or confidence/validation fails, use approved extraction or manual review.
Create a native invoice only after the displayed evidence is complete and then verify the stored source file hash.
Fields and decisions to review
This guide’s focusDetailed client guidance covering process a machine-readable PDF invoice with Sage Document Intake, including prerequisites, exact controls, verification evidence, failure boundaries and safe recovery steps for the shipped Sage connector.
Secondary workflowDocument Intake is a controlled migration/external capture route. Normal native invoice creation remains the preferred operating workflow.
File controlsPDFs are checked for permitted MIME/signature, file size, hashes and duplicates and are stored under randomized protected paths.
Text extractionMachine-readable PDFs can use conservative local text extraction. Scans without usable text need an approved structured extractor or manual review.
Extractor securityA configured extractor must use HTTPS on port 443, resolve only to public IP space, verify TLS, avoid redirects and embedded credentials, and remain pinned to the resolved host.
Creation gateAutomatic native invoice creation is off by default and uses a 99% confidence default when enabled. Summary-line fallback is also off by default.
ValidationClient match, invoice dates, currency, totals, tax and line reconciliation must pass before creation can be considered safe.
How to confirm it worked
Reopen the native record and the relevant integration history/register. Confirm the stored remote accounting identifier, exact amount/status, latest timestamps and the expected audit or reconciliation evidence. If the provider-side state cannot be independently verified, keep the task unresolved and use the controlled exception or troubleshooting flow.
Questions clients commonly ask
Can every scanned PDF be created automatically?No. A scan with no trustworthy machine-readable structure requires an approved structured extractor or manual review; auto-create remains off by default.
What should I verify before I trust the result?Verify the native record, the mapped remote record, the stored remote identifier, the current status/amount and the audit or history evidence. A success message alone is not the accounting evidence.
Should I press the action again after a timeout?Not until you have checked the remote accounting service and the Exceptions/API evidence. A timeout can occur after the remote side accepted a non-idempotent write, so a blind retry can create a duplicate.
Does the connector replace the native CRM record?No. Native clients, invoices and payments remain the operational source records. The connector records mappings, remote evidence and controlled reconciliation around them.
What should I include in a support case?Include the native record ID, remote ID if known, date/time, action, exact visible error and request/audit identifier. Never send client secrets, access tokens, refresh tokens or authorization headers.
Technical basis for this guidance
This guide was checked against the supplied module code paths: controllers/Sage_accounting_automation.php; models/Sage_accounting_automation_model.php; libraries/Sage_api_client.php; libraries/Sage_reconciliation_service.php; libraries/Sage_document_import_service.php; views/admin; install.php; docs/FEATURE_MATRIX.md. It documents shipped behaviour, validation, routes and evidence controls; it does not claim successful live provider network calls from offline inspection.
Controls, checks and common mistakes
Work in the intended tenant and remote accounting/business context.
Do not paste credentials, access tokens, refresh tokens or private keys into tickets or notes.
Do not bypass a missing mapping, validation error, permission gate, duplicate check or remote-verification requirement.
Do not repeat a non-idempotent write until you know whether the previous request reached the remote service.
Keep native record, remote evidence and audit/history state aligned before closing the task.
Provider and accounting boundaryThe integration fails closed when required mapping, remote identity, validation, concurrency or verification evidence is missing. Do not bypass a blocker by editing unrelated records or repeating an uncertain write.
