The Public Service in 2035

Whenever automation in public administration comes up, the same reflex question follows: how many jobs will disappear? It is the wrong question. Over the next ten years, the public service will not have too many people but too few, and the self-driving state answers the question of how fewer staff can deliver more and better services. This article describes which roles will emerge by 2035, how leadership changes, and why public administration gains as an employer as a result.

The math that does not work without software

In Austria, around 45 percent of federal staff will retire by 2034; in the IT departments of the federal administration, this wave arrives even earlier (Die Presse 2022, cited in "Der selbstfahrende Staat", Springer Gabler 2024). These baby boomers take deep expertise and decades of experience with them. Replacing them one for one is arithmetically impossible: even if every graduate of a relevant IT program joined the civil service, their number would fall far short. The figures differ from country to country; the direction does not.

At the same time, the volume of data that public administration has to handle is growing exponentially, not linearly, and with it the flood of inquiries. Anyone who wants to close this gap with the means of the analog state has only two options: recruit ever more staff who do not exist, or lower the depth of review and put quality, legal certainty, and revenue at risk. The example of Vienna's Municipal Department 35, on which the Austrian broadcaster ORF reported on August 17, 2021, shows where the second option leads: phones that ring all day and go unanswered because every answered question triggers further inquiries.

The debate about job cuts therefore misses the point. Software does not replace a position that remains vacant anyway. The real question is: how can an administration that will have to manage with significantly fewer people in 2035 fulfill its tasks better than it does today? That is precisely the question the self-driving state answers.

What software handles in 2035, and what it does not

In our book, we assume that in the self-driving state around 80 percent of all routine work is handled by software: applications whose requirements are evidenced by registers are decided without human intervention, deadlines are monitored, payments are triggered, and notices are delivered. The remaining 20 percent or so are the cases that require discretion, judgment, or empathy. For these, staff have an up-to-date, cross-departmental data basis at their disposal: decision variants can be simulated, and the fairest variant is approved and triggers the administrative act. We described how registers replace paper evidence in the article Registers Instead of Certificates.

One burden almost disappears entirely: documentation. A large share of today's work consists of justifying decisions in writing, recording them, and filing them. In the self-driving state, every transaction is automatically certified and traceable; the justification is created with the decision, not after it. How this is reconciled with the requirements of the rule of law was the subject of the article Explainability as a Principle of the Rule of Law.

This also changes the timeline of major projects. The building permit for converting a former industrial site into a residential quarter with all its infrastructure, today a procedure that takes years, can be completed in a few weeks, provided the political decision has been made and the organizational units involved work together in an agile manner. Not because less is examined, but because gathering, coordinating, and recording are no longer the main work.

New roles instead of job cuts

What is left for people when routine work disappears? More than the debate suggests. In the book, we describe fields of employment that emerge or grow in the self-driving state. The most important ones for the public service can be summarized in a comparison:

Three of these roles deserve a closer look.

Advice and assistance. For people who cannot use digital services, or cannot use them fully, personal support becomes more important, not less. Administrative staff in service centers answer questions, give advice, and, on request, visit people at home or at their business. The administration comes to the person, not the person to the administration. What stands out in these conversations flows back to the specialist departments and political bodies as suggestions for improvement.

Quality management of decisions. If software decides 80 percent of the cases, people are needed who continuously review these decisions, make them understandable, and refine the rules. This is not a residual task but the core competence of an administration that takes responsibility for its algorithms. Anyone who keeps case files today knows the cases where rules fail and is better qualified for this role than most job descriptions would suggest.

Administrative informatics. Software developers, data analysts, and security experts will remain scarce in the public service in the medium term. That is precisely why administrative informatics, the combination of legal knowledge and IT, is the role into which experienced caseworkers can develop. They know what a law means; when translating it into software, that is the harder half.

Added to this are the empathetic social professions in social work, care, education, and migration, where demand is rising because individually tailored offers are available for every single case and there is time for them. One of the six theses on the role of humans, which we first formulated for the self-driving company ("The Self-Driving Company", Springer 2023) and which comes second in the book on the self-driving state, sums it up:

Humans will always have work.

This is not a reassurance but an assignment for education systems, HR departments, and every individual to build the skills that will be needed in 2035.

Leadership: from the three C's to support, challenge, and feedback

Leadership changes along with the roles. Traditional administrative management rests on three C's: command, control, and correct. When software takes over administrative control, an approach with three different foci takes its place: support, challenge, and give feedback. Managers are relieved of control tasks and can devote themselves to what only people can do: motivation, personal support, team spirit, and enthusiasm for meaningful tasks.

In the book, we describe two team structures for this. Self-organizing teams pull their tasks independently from a backlog and handle the cases that require creativity and empathy, such as individual coaching in job placement. Software-driven teams perform clearly defined tasks according to precisely prepared work instructions, for example in infrastructure maintenance with mobile workforce management. Both forms are steered by goals, not by instructions; objectives and key results (OKR, Kudernatsch 2021) are a suitable instrument for this because they set direction and leave autonomy.

Two consequences matter for workforce planning. First, teams can grow larger when the administrative burden disappears; experience from industry, which we refer to in the book, shows that spans of control can be significantly widened with technical support without the quality of leadership suffering. Second, management positions no longer necessarily have to be full-time positions. This opens management careers to people who currently drop out because of caregiving responsibilities or part-time work, and it is one of the underestimated levers for the attractiveness of the public service.

Attractiveness as an employer

In 2035, the public service will compete with every other employer for the same scarce professionals. I am convinced that it can win this competition, but not with today's arguments. In my observation, job security and pension entitlements attract young people less than they used to; what attracts them is the content of the work. The self-driving state changes exactly this content:

• Purpose instead of Sisyphus. Those who hand routine work over to software work on the cases that count and in direct contact with people. In the book, we describe Klaus Stadtmeister, who works in a city administration and, instead of paperwork, develops the redesign of a park together with the residents. That is not a utopia but the description of a job that can exist in 2035.
• Tools that are up to date. An administration in which data is immediately available and variants are simulated with a click is a more attractive workplace than one in which evaluations are compiled by hand from spreadsheets.
• Flexible working models. When leadership no longer has to mean full-time and teams organize themselves, part-time work, role changes, and re-entry become the norm.
• Better pay where it counts. The self-driving state frees up funds through saved administrative costs that can flow into the pay of professions with staff shortages. In the book, we describe this feedback loop using the example of nursing care: professions become more attractive, young talent is trained in a more targeted manner, and people stay in the profession longer.
• Further qualification by interest. Since many routines disappear, existing staff can be further qualified according to their interests and abilities instead of being pushed into residual tasks.

The image of the state as an employer thus turns around: from the authority in which one is administered to the organization in which one shapes things.

What administrations can do now

The retirement wave will not wait for 2035. Three steps can be started today:

1. Plan from the target picture. Workforce planning that only refills positions reproduces yesterday's administration. Those who know which roles will be needed in 2035 can use every vacancy as an opportunity to change the profile instead of copying it.
2. Secure the knowledge of those leaving. The baby boomers take experience with them that appears in no procedural documentation: which cases are difficult and why. This knowledge belongs in the rules of the software before the people leave. Automation therefore begins with listening.
3. Make further qualification a permanent state. The role of quality management for automated decisions or of administrative informatics does not come about through job postings but through developing the people who are already there. Managers bear the responsibility for this; leadership through support, challenge, and feedback begins right here.

The next step

The roles, the team structures, and the six theses on humans in the self-driving state are described in detail in the book; the page The Self-Driving State provides an overview. How we accompany administrations on the way there, from strategy to implementation, is described on the page Public Administration. If you want to align your organization's workforce and role planning with this target picture, we would be glad to talk with you about it.