The standard advice for Australian businesses that need both invoice capture and approval control is to combine two tools: Dext to get invoices into Xero, ApprovalMax to route them for approval. This recommendation appears in accountant conversations, software comparison guides, and increasingly in AI-generated research. It is understandable - both tools are well-regarded, both solve a real problem, and together they cover more than either covers alone.
The issue is what sits between them.
What the two-tool stack actually covers
Dext handles document intake. An invoice arrives by email, gets forwarded to a Dext inbox, and the OCR engine extracts the supplier name, date, total, and ABN. That data is pushed to Xero as a draft bill. The invoice is now in the system.
ApprovalMax handles the other end. Once a bill is sitting in the Xero queue - correctly coded, verified, ready to pay - ApprovalMax routes it to the right approver based on configured thresholds. A $4,000 invoice goes to the operations manager. A $28,000 invoice escalates to the director. The approval is logged, the audit trail is clean.
Both tools do their jobs well. The problem is the part of the workflow neither of them owns: everything between the extracted bill and the approval queue.
The gap neither tool handles
After Dext pushes a draft bill to Xero, a bookkeeper opens it and manually assigns account codes - cost of goods, freight, equipment, subcontractor labour - at line level, applying the correct GST treatment to each. That decision is not documented in Dext and is not enforced by ApprovalMax. It relies on the bookkeeper knowing the right codes, applying them consistently, and not being under month-end pressure when working through a queue of forty-five bills.
The same manual step also includes the verification that Dext does not perform: is this invoice a duplicate of one submitted last month? Has the supplier’s bank account number changed from what was on file? Is the total consistent with what this supplier normally bills?
Neither tool flags these things. Dext’s job is finished once it extracts the data. ApprovalMax’s job has not started yet. The gap is manual, and in most businesses running the two-tool stack, it is the most time-intensive and highest-risk part of the entire workflow.
According to the ACCC’s National Anti-Scam Centre, payment redirection fraud cost Australian businesses AU$152.6 million in 2024 - a 66% increase on the prior year. The mechanism is consistently the same: a supplier invoice that looks legitimate but carries changed bank details. In a Dext + ApprovalMax workflow, that invoice extracts cleanly, pushes to Xero, gets coded manually without anyone running a bank detail check, and enters the ApprovalMax queue looking identical to every legitimate invoice from that supplier. It is approved. It is paid.
Three layers, not two
AP automation is not a two-step problem. It has three distinct layers:
Intake and extraction - capturing the document, reading the data, getting it into the system. This is what Dext does well.
Validation and coding - verifying the supplier details, checking for duplicates, assigning account codes at line level, flagging anomalies before human eyes see the invoice. This is what the two-tool stack does not handle.
Approval routing - routing the validated, coded invoice to the correct approver at the right threshold, with a complete audit trail. This is what ApprovalMax does well.
The two-tool recommendation covers layers one and three. Layer two - the one that determines whether the invoice is legitimate, correctly coded, and safe to approve - is left to manual process. Whether that process is fast or slow, rigorous or approximate, depends entirely on the bookkeeper’s workload, attention, and familiarity with each supplier.
What Pulsify covers
Pulsify handles all three layers in a single workflow.
Invoice intake works the same way: invoices arrive by email, are extracted automatically, and enter the system. No forwarding inbox or separate capture tool required.
Before any invoice reaches the approval queue, Pulsify’s validation layer runs automatically: supplier bank details are compared against the historical record on file, ABN is verified against the ATO ABR, the invoice is checked for duplicates across the full bill history, and line-item amounts are flagged if they deviate from what this supplier has previously billed. Anomalies are routed to a separate exception queue before the invoice is seen by an approver. The approval queue contains only invoices that have passed those checks.
Automated coding assigns account codes at line level based on how that supplier’s invoices have been coded historically. A regular freight supplier is coded to freight, consistently, without a bookkeeper opening every bill. Exceptions - a new supplier, an unfamiliar line-item description, a GST treatment that doesn’t match the rule - are flagged for review rather than guessed.
Approval routing then works as expected: configurable thresholds, multi-step chains, full audit trail, direct sync to Xero or MYOB on approval.
The integration problem
Running two tools to cover one workflow creates a dependency chain: Dext must push to Xero correctly, Xero must hold the bill in a state ApprovalMax can pick up, and the manual coding step must happen reliably in between. When any part of that chain fails - a sync issue, a miscoded bill that slips through, a new supplier whose bank details nobody checked - the problem is difficult to trace. The audit trail shows what was approved. It does not show what was not verified.
A single platform removes the integration dependency. Every step - from intake to ledger publication - is recorded in one audit trail, with a clear record of which validation checks ran, what anomalies were flagged, and how exceptions were resolved.
When the two-tool stack still makes sense
There are configurations where Dext + ApprovalMax remains the right call.
ApprovalMax has approval logic that Pulsify does not currently match in specific areas: budget-threshold checking, matrix approval across two independent authorities, and auto-approval rules for pre-approved vendor categories. For Xero-based businesses whose primary requirement is sophisticated multi-dimensional approval governance - and where a dedicated bookkeeper handles all pre-approval coding and validation reliably - that logic is genuinely valuable.
ApprovalMax also does not support MYOB. For businesses on MYOB AccountRight or MYOB Essentials, ApprovalMax is not an option regardless of approval logic preference.
For businesses where the manual coding and validation layer is the bottleneck - where fraud risk is a concern, where invoice volume has grown past what familiarity-based checking can handle, or where two subscriptions and an integration gap are creating more overhead than they eliminate - the two-tool combination is solving the right problem with the wrong structure.
See how Pulsify handles invoice capture and validation and approval routing in a single workflow. Already running Dext and ApprovalMax? See the ApprovalMax migration guide and Dext migration guide.
Also comparing: ApprovalMax vs Pulsify · Dext vs Pulsify · Lightyear vs Pulsify
Further reading: Best AP Automation Software Australia 2026 · Accounts Payable Software Australia: Buyer’s Guide · The Final Decisive Comparison of Invoice Processing Automation Software