A child is born in Estonia. The parents do not begin their first exhausted week by hunting for the correct family-benefit form, proving facts that the state already knows and carrying the same information from one office to another.
Once the birth is entered in the population register, authorised systems can check relevant information—including income and employment data—and calculate which family benefits may apply. The parents receive an electronic offer explaining the entitlement and the calculation. They still confirm details such as the bank account and decide whether to accept. The service has begun because something happened in their lives, not because they discovered the right administrative door.
This is the promise of event-driven government: public services organised around birth, unemployment, retirement, bereavement, moving home or starting a company, rather than around ministries and their forms.
It sounds obvious. It was technically imaginable decades ago. So why has government usually forced the citizen to start every procedure?
Why government waited for us to knock
Traditional administration was built as a collection of institutions, each created by a different law, with its own budget, vocabulary, records and responsibilities. A birth belongs to the civil registry. Family allowance belongs to social insurance. Childcare may belong to a municipality. Health coverage sits somewhere else again.
Paper made these boundaries physical. Early computerisation often reproduced them: the filing cabinet became a database, but it remained a departmental database. Putting a PDF form on a website did not change the underlying organisation.
There were also good reasons for caution. A civil registry should not casually broadcast a birth to every public body. Data-protection law limits the purposes for which information may be reused. Records can be incomplete or contradictory. One agency may record a “household” differently from another. And public officials need a clear legal mandate before initiating a payment or making a decision.
Finally, the incentives were wrong. The cost of administrative fragmentation fell largely on the citizen, while the cost and risk of integration fell on the agencies. No individual department owned the whole experience.
What changed?
No single invention produced proactive services. Several foundations matured at the same time.
- Reliable digital registers made births, addresses, companies and entitlements available as structured, machine-readable records.
- Strong digital identity and electronic signatures made it possible to identify a person remotely and obtain a legally meaningful confirmation.
- Secure data-exchange layers and APIs allowed agencies to request specific facts without assembling one gigantic central database.
- The “once-only” principle gained political and legal support: citizens should not repeatedly submit evidence already held by an authorised public body.
- Workflow engines, rules systems and event-streaming technology became mature enough to coordinate long-running processes across old and new systems.
- Cloud infrastructure, observability and cybersecurity controls made it easier to operate services continuously and record who accessed which data.
Estonia’s X-tee data-exchange layer, based on the open-source X-Road technology, illustrates the model. Each institution retains its own system, but authorised exchanges are authenticated, encrypted, signed and logged. The architecture connects registers without pretending that all government data should live in a single pot.
Policy changed too. The European “once-only” agenda and projects in Estonia, Lithuania and elsewhere explicitly require services to be redesigned around events. Lithuania, for example, has combined services for childbirth, employment changes, moving country and starting farming. Kazakhstan links birth registration with childcare benefits and kindergarten enrolment. Norway and Slovakia can provide certain childbirth-related payments without a separate family application. These examples differ in how far they automate the final decision, but they share a direction: the administration notices a relevant event and offers the next useful step.
What does the machinery look like?
An event-driven service is not simply an AI chatbot. Its core is usually more prosaic—and more dependable.
- An authoritative register records an event: a birth, death, job loss, new address or company registration.
- An event broker or secure integration layer tells only authorised systems that something relevant occurred.
- A rules engine checks legislation and eligibility criteria.
- A process orchestrator coordinates requests to registers, handles missing information, waits for responses and routes exceptional cases to a person.
- An identity and consent layer lets the citizen authenticate, review the proposed action and provide information the state does not possess.
- A case-management system records the decision, reasons, deadlines and any appeal.
- A notification service reaches the citizen through a secure inbox, app, email, letter or—where appropriate—SMS.
Artificial intelligence can help interpret unstructured documents, translate explanations or guide a caseworker. But entitlement rules should normally remain explicit, testable and auditable. A probabilistic language model is a poor substitute for a clear legal rule.
Who supplies the technology?
There is no single “event-driven government platform.” Administrations combine public digital infrastructure, open standards and commercial components.
At the public and open-source end, X-Road provides secure data exchange; GovStack defines reusable government building blocks; and systems such as MOSIP provide modular digital-identity infrastructure. The EU’s Once-Only Technical System and emerging European Digital Identity Wallet add cross-border evidence exchange and verifiable credentials.
For process orchestration, governments may use products such as Camunda or open BPMN-compatible engines. Event distribution can use technologies such as Apache Kafka and commercial platforms built around it. API-management and integration products connect legacy registers that cannot be replaced overnight.
Large vendors including ServiceNow, Salesforce, Microsoft, Oracle, SAP, AWS and Google offer combinations of government cloud, case management, identity, workflow, data integration and AI. Specialist national contractors often perform the less visible but essential work of mapping legislation, cleaning registers and connecting old systems.
The list should not be read as an endorsement. The best architecture is deliberately modular. A government should be able to replace its notification tool, workflow engine or AI model without losing control of the rules, records and public-service process.
The darker version of invisible government
A service that arrives at the right moment can feel humane. The same infrastructure can also create an administration that observes too much and acts too quickly.
Bad data can become automated injustice. A wrong address, income record or family relationship may now block several services at once. Efficiency for the majority can make the exceptional case harder to see.
Proactivity can become surveillance. Combining data to offer childcare is not equivalent to combining it to score a family’s behaviour. Purpose limitation, data minimisation and strict access controls are therefore not administrative obstacles; they are part of the service.
An invisible decision is difficult to contest. If no application was filed, citizens may not even realise that an algorithm evaluated them and found them ineligible. A responsible system should therefore communicate negative or incomplete outcomes, not only successful ones.
Digital exclusion remains real. A smartphone cannot be the price of citizenship. Event-driven services need assisted, telephone and paper routes, especially for people facing disability, language barriers, homelessness or domestic instability.
Integration increases the blast radius. A compromised identity service or integration layer can affect many agencies. Shared infrastructure needs unusually strong security, independent audits and resilient offline procedures.
Vendor lock-in can become constitutional in scale. When a private platform encodes eligibility rules and coordinates essential services, changing supplier is no longer an ordinary IT migration. Open standards, data portability and public ownership of process definitions matter.
Does the citizen remain in control?
Partly—but “consent” is not a complete answer.
In Europe, a public administration often processes data because legislation authorises a public task, not because each citizen clicks “I agree.” A person cannot withdraw consent from taxation or population registration. Yet citizens retain important rights of access, correction, objection and, in significant cases, protection against solely automated decisions. They should be able to request human intervention and contest a decision. The European Commission’s guidance on automated decision-making explains these safeguards.
Control should therefore be designed at several levels:
- a dashboard showing which authority accessed which information and why;
- the ability to correct source data, not merely the final form;
- a clear explanation of the rule and evidence behind a decision;
- human review and an accessible appeal;
- choice over optional data sharing and communication channels;
- an opt-out where automation is convenient rather than legally necessary;
- independent oversight, published impact assessments and meaningful statistics on errors.
The European Digital Identity Wallet points toward a useful model: credentials stay under the user’s control, selective disclosure limits what is revealed, and a privacy dashboard records sharing. But a wallet alone cannot govern data exchanges that take place under statutory authority behind the scenes.
How citizens can help
Citizens do have influence, although it is mainly political and collective rather than a settings button.
They can participate in consultations and user-testing panels, support civic-technology organisations and data-protection groups, and ask elected representatives concrete questions: Is the service optional? Which registers are consulted? Are unsuccessful eligibility checks disclosed? Can a person see the access log? Who audits the code? What happens when the data is wrong?
People can also use well-designed digital services, report confusing journeys and keep core records accurate. Adoption gives governments evidence that integration is worthwhile. But citizens should resist the idea that convenience requires a blank cheque. Supporting proactive services and demanding strict limits are not contradictory positions; together they create trustworthy infrastructure.
The democratic bargain should be simple: if the state is allowed to become more perceptive, it must also become more visible, correctable and accountable.
The next challenge: from isolated successes to a public-service fabric
The technology for a handful of event-driven services already exists. The harder future challenge is scaling it without creating a universal surveillance machine or a brittle mega-platform.
Cross-border life events will test Europe in particular. Moving from Germany to France, studying abroad or caring for a parent in another member state still exposes the limits of national registers and legal vocabularies. Digital identity wallets and once-only evidence exchange may reduce the friction, but technical interoperability is only half the problem. Laws must agree on what a credential means, which authority may request it and how an error is corrected across borders.
Governments must also learn to recognise what automation should not do. Birth registration and a universal child benefit are relatively clear. Disability, vulnerability, family risk and unemployment sanctions involve context and judgment. The future of GovTech will not be measured by the number of decisions made without humans. It will be measured by whether technology removes needless effort while preserving dignity, discretion and the possibility of being heard.
The most advanced public service may eventually be the one the citizen never had to search for. The most democratic one will still leave a visible door through which the citizen can walk back in.
Sources and further reading
- OECD: Modernising Access to Social Protection
- OECD: Personalised and proactive public services
- European Commission: Event services in Estonia
- European Commission: The once-only principle
- EU Digital Identity Wallet: Security and privacy