Short answer: this no longer looks like science fiction. Ukraine already has a mature digital-government ecosystem, tens of millions of Diia users and a functioning Diia.AI assistant that is moving from answering questions toward completing real public services. But there is still a large gap between “AI exists in Diia” and “AI is the main interface between citizens and the state.” The current probability of that transition by the end of 2030 is estimated at 60%, with a working range of 55–65%.
Forecast details
Probability: 60%
Confidence: 74/100
Horizon: December 31, 2030
Snapshot: October 14, 2026
Base Rate: N/A — the technology is too new for a reliable historical frequency
Historical Similarity: N/A — international peers are themselves still in transition
Here, “interaction with the state” primarily means high-volume digital public services: certificates, benefits, documents, registrations, transport and social services, entrepreneurship and access to government information. It does not mean AI replacing courts, police, doctors, military service or every in-person procedure.
What would count as YES
The forecast resolves RESOLVED — YES if, by December 31, 2030, the Diia AI assistant or its successor becomes the main entry point for digital public services. Users should be able to find, start and complete most high-volume services through text or voice without manually searching menus, and official data or a representative national survey should show that the AI interface is used by a majority of citizens interacting with digital public services.
If Diia.AI remains popular but most services are still completed mainly through menus, forms and separate service pages, the forecast resolves RESOLVED — NO.
Ukraine already has the foundation
According to the Ministry of Digital Transformation, Diia has more than 24 million app users, more than 7.5 million portal users and over 190 services. A UNDP-backed study with KIIS found that 59% of Ukrainians used government e-services in 2025, with Diia remaining the leading channel, while 85% of users reported a positive experience.
This gives Ukraine an important advantage. AI does not have to move citizens directly from paper bureaucracy to conversations with an algorithm. It can become the next interface on top of an already familiar digital state.
Diia.AI has already moved beyond a simple chatbot
The first Diia.AI version launched on the web portal in September 2025, followed by the mobile assistant in May 2026. Users could already obtain selected extracts, pay traffic fines and generate an income certificate, while the assistant could identify relevant services from a description of the user’s life situation. Diia described the model in its official launch announcement.
This changes interface logic. Instead of knowing the government structure or exact service name, a user describes the desired outcome in natural language and the system maps the request to the relevant process.
The strongest signal is that the government has already set this objective
In 2026, the Ministry of Digital Transformation explicitly tasked its AI leadership with turning Diia.AI into a single window for interaction between people and the state, expanding services, voice interaction and proactive assistance. The broader direction is a transition from a Digital State toward an Agentic State by 2030.
The ministry has also reported roughly 350,000 Diia.AI testers and about 2 million messages. That is real-world adoption, but still early relative to Diia’s full user base.
International experience lowers confidence slightly
Estonia is developing Bürokratt, a network of public-sector AI assistants designed to let citizens request information and services conversationally. Yet the project has been under development for years and cross-agency agent infrastructure is still being built.
The UK shows a similar pattern. Government Digital Service tested GOV.UK Chat for around 18 months, ran public pilots and only in 2026 began opening the system more broadly. Its team published a testing report.
The main barrier is legal and technical reliability
By 2030, AI will almost certainly be much better at understanding natural language. That is not enough for government. A state agent must reliably identify intent, use government data correctly, explain consequences, obtain consent, authenticate users, work with digital signatures, log actions and allow mistakes to be challenged.
The most plausible architecture is therefore AI as a conversational layer on top of deterministic government systems rather than a model freely deciding what to do with a citizen’s data.
The state cannot become only an AI chat
Even if Diia.AI becomes the easiest channel, alternatives will remain necessary. The UNDP/KIIS study shows a persistent digital divide, especially among older citizens. Physical service centres, phone support and conventional web forms therefore remain necessary.
The forecast is not that Ukrainians will interact with government only through AI, but that AI may become the default first interface while classic channels remain available as fallbacks or specialist routes.
Four scenarios through 2030
| Scenario | Probability | What it would mean |
|---|---|---|
| AI becomes the main interface | 60% | Diia.AI becomes the main entry point; most popular services can be started or completed through text or voice |
| Hybrid Diia | 27% | AI is widely used for discovery and some transactions, while menus and forms remain equally important |
| AI remains an assistant | 10% | The assistant handles guidance and simple operations, while legally significant services remain mostly in conventional interfaces |
| Development slows materially | 3% | Security, funding, regulation or technology keep AI as a secondary feature |
What would change the forecast
Probability would rise if, during 2027–2028, Diia.AI gains a large share of the most frequently used services and official statistics show a clear shift from menu navigation toward conversational interaction. Evidence that more than 30% of digital government interactions already begin through AI well before 2030 would be an especially strong signal.
The estimate would fall if Diia.AI remains limited to a small set of simple operations, fails to become a mass channel by 2028, or experiences a major data-protection or legally significant error that forces restrictions on agentic features.
Why the estimate is 60%
Ukraine already has mass digital adoption, connected registries, digital identity, electronic signatures and an AI assistant capable of real operations. Most importantly, the government has explicitly chosen the Agentic State direction.
But scaling from a handful of operations to the default interface for millions of citizens is much harder than launching a chatbot. Legal correctness, security, accessibility and trust all matter, and international experience suggests that even digitally mature states are moving cautiously.
Forecast History: October 14, 2026 — Probability 60%, Confidence 74/100. Initial snapshot.
This forecast does not claim that the event will happen. It reflects the current estimated probability based on information available at the snapshot date. The estimate may change as new evidence appears.
