The Office Without Forms

The form is the most visible piece of public administration: it stands at the beginning of almost every procedure and marks the moment when the state asks citizens to tell it what it usually already knows. In the self-driving state, the form disappears, but not because it becomes digital. It disappears because the procedure runs itself: it starts with an event, draws its data from the registers, and decides according to a stored set of rules. This article describes what a procedure needs in order to do that, where it already works today, and what public administrations can take from it for their current digitalization projects.

The form is a symptom, not the problem

In the first two articles of this series, we explained why public administration is the ideal candidate for the self-driving principle and described the four dimensions along which the state evolves. Today we turn to the individual procedure, the place where it is decided whether digitalization actually changes administration or only its surface.

Anyone who converts a paper form into an online form has not yet changed the administration. Citizens type on a screen what they used to write by hand; supporting documents are uploaded as PDFs instead of photocopied; the review is still done by a person who checks the entries against the registers. That is better than paper, but it does not change the logic: the state asks for data it already holds.

Hendrik Scholta, Willem Mertens, Marek Kowalkiewicz, and Jörg Becker described this logic precisely in 2019 in their stage model "From one-stop shop to no-stop shop." They distinguish three dimensions of digital maturity: the integration of data collection (from a separate form for every request to no form at all, because all the information is already known), the integration of data storage (from departmental solutions to government-wide data management), and the purpose of data use (from reactive to proactive and predictive). At the end of this development stands the no-stop shop: services are delivered without anyone having to stop and file an application. In our book Der selbstfahrende Staat (Springer Gabler 2024, in German), we describe the no-stop shop as the target picture that is embedded in the DNA of the self-driving state.

In this model, the form is a symptom: it indicates that data collection, data storage, and data use are not yet integrated. Anyone who wants to abolish it must change not the form but the procedure behind it.

Four stages of a procedure, using unemployment benefits as an example

In the book, we describe every life situation along four evolutionary stages: analog, digital, automated, and self-driving. The life situation of unemployment shows clearly how the initiative shifts from citizens to the state along the way.

The decisive transition lies between the third and the fourth stage. In the automated state, the application is still the trigger; only the processing is software. In the self-driving state, the event is the trigger: the end of an employment relationship is known to the state through the deregistration from social insurance; if no new registration follows, the procedure starts ex officio. The form has lost its function because the information it was supposed to carry is already in the system.

The same pattern appears in other life situations. At a birth, the hospital administration records the child's data, the birth certificate is created automatically in line with the once-only principle, and the follow-up processes, that is, notifications to authorities, identity documents, and benefits, start without any further application. As soon as the state knows about a pregnancy, it can itself trigger the invitation to the mother-and-child consultation and the family allowance. And in emergencies such as a pandemic, aid can be calculated and distributed according to need from the existing data pools, instead of reviewing physically submitted applications.

What already works without an application today

That sounds like 2040, but part of it has been reality in Austria since 2017: the application-free employee tax assessment. Employees who do not file a tax return still receive their tax refund in many cases, because the tax office prepares the assessment from the data it holds anyway: income, social security contributions, and payroll tax deductions. The credit is transferred without an application. This relieves above all those employees who are less familiar with tax matters, and it spares everyone concerned the collecting of documents and the filling in of forms.

Two things about this example matter for practice. First, in 2024 there is no sophisticated artificial intelligence behind the application-free assessment, but the mathematical calculation of tax amounts according to a complex yet formal set of rules. Second, the prerequisite was not a new technology but the decision to start the procedure ex officio and to use the existing data for it. Application-free administration is first a question of data and rules, and only then a question of software.

In the book, we assume that already at the stage of automated service delivery, presumably 50 to 60 percent of administrative decisions can be made automatically, because they are defined so unambiguously by laws and regulations that they can be represented as a calculation or a decision chain. We describe the automated state as a whole as one in which around 80 percent of processes run on the basis of software algorithms and people make the critical decisions: whether a four-eyes principle is required, whether an expert opinion needs to be obtained, whether further inquiries are necessary.

Four prerequisites for a procedure that runs itself

From the building blocks we describe in the book for self-driving service delivery, four prerequisites can be derived that a procedure must meet in order to do without a form.

1. A triggering event. Every procedure needs a beginning. In the state of forms, that is the application; in the self-driving state, it is an event the state already knows about: a birth in a hospital, a deregistration by the employer, a change of address, the expiry of a deadline. The first design question is therefore not "What does the form look like?" but "Which office is the first to learn of the life situation, and how does this information get into the procedure?"
2. Registers that count as the source. The once-only principle, which the EU anchored in Regulation (EU) 2018/1724 establishing a single digital gateway, means that data is recorded once at the responsible office and then used by all authorized offices. Whether the registers are kept centrally or federally is secondary; both models are equivalent in the book. What matters is that for every data item it is clear who maintains it, and that mutual access is regulated and logged. The residents' registration system shows how closely this is tied to organization: the central register of residents is national, the registration itself is organized through the states and municipalities, and for registration via an app such as "Digitales Amt" (Digital Office) to work in all municipalities, the federal level needs interfaces to all of them.
3. An executable set of decision rules. Software can only decide what exists as a rule. In the book, we call this set of rules the digital legal twin: the set of rules and explanations that define how a decision is made algorithmically, complemented by algorithmic auditing procedures that ensure the software works correctly. For the application-free employee tax assessment, such a set of rules has long existed; for many other services, it would first have to be extracted from statutes, regulations, and decrees.
4. An interaction protocol instead of a file. The paper file has had its day. It is replaced by a continuously updated protocol in which input data, decision bases, the steps carried out, and the decision itself are recorded traceably. If, for example, the social insurance institution opens an assessment procedure, it no longer requests supporting documents but accesses the required data ex officio and logs the access. Only in exceptional cases is data missing; then exactly that data is requested, not the whole form.

The form does not disappear because it becomes digital, but because the procedure starts on its own.

Who decides, and who stays at the counter

An office without forms is not an office without people. In the book, we describe the Public Service Wallet, in which citizens have their documents, identity cards, and authorizations at hand at all times and can proactively grant approvals or reject data requests. Linked to it is an entitlement: to receive state services without having to apply for them. Conversely, there is a kind of basic right to receive them without owning any technical devices. Anyone who does not want to use a smartphone goes to a regional service center or turns to a trained person who takes on the request and operates the systems. For the software behind it, it makes no difference by which route the request arrives. On request, every form remains available in analog form, for instance when a passport is needed for a country that does not yet accept digital documents.

For caseworkers, the work shifts. Instead of receiving, checking, and entering the same forms, they look after the people who come in person, handle the special cases for which the rules have no answer, and make the decisions that require a four-eyes principle or an expert opinion. This requires requalification, which we describe in the book as one of the first steps on the road to the self-driving state. Anyone planning digitalization projects today should treat this part not as an accompanying measure but as the core of the undertaking.

What public administrations can take from this today

In our project practice in the public sector, we see that many digitalization projects begin with the form and end there as well. Five questions help to think a procedure from the event instead:

• Which event triggers the procedure, and who learns of it first? If an office within the state already knows the life situation, the application is superfluous.
• Which fields in the form does the state already know? For every field, it should be stated in which register the value is held. What lies in a register does not belong in a form.
• Which decision rules are embedded in the procedure, and are they written down? Rules that exist only in the heads of experienced caseworkers can be neither automated nor audited.
• What happens when data is missing or the result is disputed? The exceptional case needs a return channel, a human review, and a legal remedy before the standard case is automated.
• What share of services is delivered without an application? This metric measures progress more honestly than the number of forms available online.

Procedures that answer these questions can run themselves. The form is then not abolished; it becomes superfluous.

The next step

The four evolutionary stages, the service catalog of life and business situations, and the technological components of self-driving service delivery are described in detail in the book Der selbstfahrende Staat: About the book. How we support public administrations in rethinking their procedures from the event onward is shown on the page Public Administration.