Validate visa applications at lodgement instead of asking questions later
The same nomination decision takes 5 days at the median and 9 months at the 90th percentile. The difference is not case difficulty — it is whether a document was missing and the department had to ask. That is a validation problem, and validation is the one thing software is genuinely good at.
Steps this would change
Rationale
This database rarely gets to point at a single published pair of numbers and say there, that is the whole finding. Here it can.
Department of Home Affairs published processing times, August 2026:
| Step | 50th percentile | 90th percentile |
|---|---|---|
| Standard business sponsorship | 12 days | 4 months |
| Subclass 482 nomination | 5 days | 9 months |
| Core Skills stream visa | 69 days | 10 months |
| Specialist Skills stream visa | 8 days | 43 days |
Five days against nine months. A factor of fifty-four, for the same decision, by the same department, under the same law, on the same form.
Case complexity does not vary by fifty-four times. What varies is whether the application arrived complete. A nomination with a defensible genuine position statement, reconciled salary evidence and compliant labour market testing is decided without a request for further information. One with a gap generates an RFI, and the file goes back into a queue.
So the 90th percentile is not measuring assessment. It is measuring the round-trip cost of a missing document — and that cost is borne by an employer who has a vacancy, a worker who has resigned, and a department that now has to read the same file twice.
Two changes:
- Validate at lodgement. Structured checks on the things that generate most RFIs — is the labour market testing evidence within the required period, does the salary evidence reconcile, is the position statement present and addressing the required elements — refusing lodgement rather than accepting an application that will predictably stall. Better to fail in five minutes at the keyboard than in five months in a queue.
- Publish the RFI rate. By visa stream and by nomination type. A department that publishes its percentiles but not the mechanism producing the spread is publishing the symptom and withholding the diagnosis. If most of the p90 tail is RFI-driven, that is a measurable, fixable, and currently invisible problem.
Who acts: the Commonwealth alone. No legislation is required to validate a form or to publish a rate.
A caution on the arithmetic. The critical path of the sponsored visa process models at 171 days on median figures. Adding the published 90th percentiles step by step gives a much larger number, but that sum is not the 90th percentile of the whole process — percentiles do not add, and an applicant unlucky at one step is not necessarily unlucky at the next. The honest statement is narrower: the tail at each step is several times the median, the tail is what wrecks a hiring plan, and the tail is the part nobody manages.
What is being traded away
Hard validation excludes people at the door. A system that refuses lodgement until every field satisfies a rule will reject applications that are correct but unusual — an atypical employment history, a salary structure that does not reconcile in the expected way, an occupation that fits awkwardly. Today those applications generate an RFI and eventually succeed. Under strict validation they may simply not be lodgeable, and the applicant has no decision to appeal because there is no decision. Validation must therefore be advisory-plus-warning for anything requiring judgement, and hard only for genuinely mechanical checks.
It moves cost onto applicants and their agents, not out of the system. The work of producing a decision-ready application does not disappear; it is done earlier and by someone else. For a large employer with a migration agent that is efficient. For a small business sponsoring its first worker it is a barrier arriving before they have any investment in the process, and the beneficiaries of "decision-ready" requirements are consistently the applicants who can afford professional help.
Publishing RFI rates creates an incentive not to ask. An officer measured on RFI rate can avoid an RFI by refusing, or by deciding on an incomplete file. Both are worse than asking the question, and both would show up as improvement. If this is adopted, RFI rate must be published alongside refusal rate and the tribunal change rate, or it will optimise the wrong thing — and migration is already one of the two portfolios driving a tribunal caseload of 131,000.
And the streams are not equivalent. The Specialist Skills stream is fast partly because of validation and partly because it selects for high income earners with well-resourced employers. Attributing the entire gap between streams to process design would overstate what any validation reform can deliver.
Precedent
The department already publishes percentile processing times, already defines what a decision-ready application contains, and already operates the lodgement systems in which validation would occur. Every component exists.
Sources
- Department of Home Affairs, published processing times, August 2026, for standard business sponsorship, subclass 482 nomination, and the Core Skills, Specialist Skills and Labour Agreement streams
- Ministerial Direction 119, in force from 25 July 2026