Digital government has become very good at giving people places to click. It is still much less reliable at answering a simpler question: what state is my matter actually in? 

That distinction matters because an online service is not only a screen. It is a sequence of consequential state changes. A form may be drafted, submitted, transmitted, received, assessed, returned, superseded, escalated, approved or closed. Those words are not interchangeable. Yet many public-facing systems still compress that history into a case number, a generic portal status or a chatbot conversation that tells the user what probably happened rather than what the authoritative record says happened. 

When the system cannot expose its own state, the citizen becomes the reconciliation layer. People keep screenshots, save confirmation emails, reconstruct timelines, remember which reference number replaced which, explain why a second submission was not a duplicate, and resend evidence an agency may already possess. The administrative burden has not disappeared. It has been exported to the person least able to see the machinery producing it. 

Australia already has much of the policy vocabulary needed to fix this. The Australian Government’s Digital Service Standard requires services to know their users, connect services, build trust in design, monitor performance and keep services relevant. The broader Digital Experience Policy now applies standards for service, inclusion, access and performance across a growing range of public-facing services. These are strong foundations. The next step should be explicit: consequential digital services should expose a citizen-visible state ledger. 

A state ledger is not another dashboard. It is a durable event history that distinguishes what the system knows from what it is still trying to determine. At minimum, it should show the current authoritative state, the event that changed that state, the time of the change, the responsible authority or system, any failed transmission, any duplicate or supersession relationship, the next action available to the citizen, and the next action owed by government. 

The design principle is simple: if a state can affect a person’s rights, money, eligibility, deadline, service access or obligation, the person should not have to infer that state from prose. 

This becomes more important, not less, as governments adopt artificial intelligence. Australia’s current AI policy already requires agencies to publish transparency statements explaining why they use AI, where the public may interact with or be significantly affected by it, and what safeguards and monitoring are in place. That is valuable institutional transparency. But service-level transparency must go further. A language model may be excellent at summarising a conversation. It should not be asked to invent the authoritative procedural state of a claim, application or case from unstructured text when that state can be represented directly. 

There is a fundamental difference between summarising evidence and declaring state. ‘The applicant appears to have provided the requested document’ is a linguistic inference. ‘Document received at 14:03; identity check pending; assessment not started’ is a system state. The first may help a worker or user understand a file. The second must come from the system of record. 

Read also: Five-year investment strengthens Australia’s First Nations language sector

A citizen-visible ledger would also make digital services easier to repair. Suppose a payment fails after an application is lodged. Today, a portal may simply show ‘in progress’, forcing the user to guess whether the application exists, whether payment is still required, whether a deadline was preserved and whether resubmitting will create a duplicate. A proper ledger can say: application received; payment attempt failed; application retained; payment retry available until Friday; do not resubmit application. One small piece of explicit state removes an entire family of support calls, duplicate records and avoidable anxiety. 

The same architecture helps when government itself makes a mistake. If an agency supersedes a case, the old identifier should not vanish. The ledger should preserve the lineage: Case A was replaced by Case B on this date for this reason; evidence items 1–7 migrated; item 8 still requires action. If an email fails to deliver, the system should distinguish ‘notice generated’ from ‘notice delivered’. If a human officer overrides an automated recommendation, that transition should be recorded without turning the internal reasoning process into public spectacle. State transparency does not require publishing every internal note. It requires making consequential transitions legible. 

This is also an inclusion issue. The people most harmed by invisible state are often those with the least spare capacity to perform administrative archaeology: people managing disability, unstable housing, language barriers, complex family situations, irregular work, limited connectivity or several government processes at once. ‘Check the portal’ is not a complete service design if the portal cannot tell them which version of reality the agency is acting on. 

A state ledger would not solve every failure in digital government. It will not make a bad eligibility rule fair, repair an understaffed programme or substitute for human discretion. It can, however, stop a different kind of failure: the situation in which government knows enough to act on a person but the person cannot see enough to understand what government believes has happened. 

That asymmetry is becoming harder to justify. Governments increasingly expect citizens to interact through authenticated portals, connected services and digital identities. In return, citizens should receive something more meaningful than convenience. They should receive procedural legibility. 

The future of digital government should not be a more persuasive chatbot sitting in front of an opaque case-management system. It should be a service in which the authoritative state is explicit enough that the chatbot, the caseworker and the citizen are all looking at the same sequence of events. 

Make the conversation intelligent, certainly. But first make the state visible. 

Joshua W.J. Brown

+ postsBio

Joshua W.J. Brown is a Toronto-based Canadian writer, filmmaker and independent systems researcher. His work examines public memory, technology, institutions and the systems that decide what counts as evidence. He writes across public policy, digital government, cultural criticism and technical systems, with work published or forthcoming in a range of literary, professional and public-affairs outlets.