Readiness is uneven
A process can cross several agencies even when those agencies have very different ways of working. One office may be ready to use an integrated review queue. Another may need a PDF sent through an established correspondence channel. A third may rely on a clerk to record a status after a telephone call or a written response. The administrative obligation exists in all three places. The digital capability does not arrive on the same date.
There are two tempting responses. One is to wait until every agency can participate in exactly the same way. The other is to launch a service that quietly excludes the difficult parts. Neither resolves the applicant’s problem. A more useful design separates the common process from the method by which each agency participates in it. That is the purpose of tiered participation.
One record, several ways to contribute
An integrated agency can receive a referral in its queue, review the relevant material and return an outcome in the platform. Its officers have accounts and permissions suited to their responsibilities. The referral, response and supporting reasons remain attached to the application. This is the most direct mode of participation, but it is not the only legitimate way to obtain an agency’s decision.
A PDF-despatch agency can receive the material in a document that fits its existing correspondence process. The central service still records what was sent, when it was sent and what response is required. When the response returns, it is attached to the same record. Sending a PDF is not equivalent to integrating the agency, and the system should not describe it as though it were.
A clerk-updated agency can participate through a designated officer who records the status and the source of that update. The distinction between “the agency decided” and “a clerk recorded the agency’s decision” matters. The audit trail should identify the action actually performed. A record can be useful without concealing how the information reached the system.
Keep responsibility visible
Tiering is not a reason to weaken accountability. Every referral still needs a destination, a request, an expected response and a recorded outcome. The service should show the participation method so an officer knows whether to expect a platform notification, a returned document or a manual follow-up. A single shared view is valuable precisely because the underlying ways of working differ.
The routing matrix is the contract between the application category and the agencies that need to review it. Not every activity requires every referral. Sending everything everywhere creates avoidable queues and makes responsibility less clear. The catalogue and routing rules should express the actual process, with a controlled way to change those rules when policy changes.
Parallel review is useful where the rules permit it. Independent referrals can proceed together instead of waiting for a folder to pass from desk to desk. Where one approval legally or operationally depends on another, the dependency should remain explicit. Software should make the permitted route faster to operate; it should not invent authority to skip a required decision.
A clock must survive the hand-off
Statutory clocks are particularly important when work leaves the platform. The absence of an integrated account should not make elapsed time disappear. The service can record the despatch, the acknowledgement where one exists, the expected response and any authorised extension. Officers can then distinguish work that is moving from work that requires attention.
This requires agreement about the meaning of each event. Is the clock running from submission, completeness or referral? Does a request for information pause it? Who can extend it, and what reason is required? These are process decisions for the responsible authority. The implementation should expose them during scoping rather than burying a convenient interpretation in code.
Design a path toward integration
A tier should be a supported participation mode, not a permanent label attached to an agency. When an office becomes ready to use the integrated queue, its access method can change while the application history remains intact. The agency identity, referral rules and decision responsibilities should not have to be reconstructed because its staff now use the platform directly.
Training can follow that progression. An office moving from despatches to an integrated queue needs to understand the same referral obligations through a different interface. A small, role-specific set of exercises is more useful than a tour of every feature. The officer should be able to receive a referral, inspect the evidence, request information and record an outcome before the new mode is relied upon.
Acceptance testing should include each supported tier. It should also include a delayed response, an incomplete return and an update entered against the wrong application. The point is to verify that the shared record remains understandable when the work does not follow the neatest demonstration path. Multi-agency software earns its value in those ordinary complications.
Progress without a fiction of uniformity
The benefit of tiered participation is practical. A service can establish a common record and visible responsibilities while agencies adopt new tools at different speeds. The applicant does not need to understand the infrastructure of each office to know that an application is moving through a coherent process. The coordinating team does need that detail, and the system should provide it.
Digital government is delivered across institutions that do not all move at the same speed. Treating that as a design condition makes the programme more honest and the service more useful. The goal is a process that works across the agencies that exist today, with a clear path for the ones that are ready to work differently tomorrow.