The growth of AI agents, rising digital threats, and the need to modernize public employees’ technology skills will shape a large share of government technology leaders’ decisions during 2026 and the years ahead. Gartner identifies these areas among its top trends for the public sector and warns that adopting new tools won’t be enough unless agencies also change their processes, governance models, and workforce readiness.
Public-sector technology trends in 20 seconds
- Gartner places agentic AI among the technologies that can reshape government operations and services.
- Rolling it out will require governance, metrics, and process reviews, not just new AI models.
- Cybersecurity will need to adapt to AI-powered threats and prepare for the shift to post-quantum cryptography.
- 78% of surveyed public-sector leaders plan to invest in employee training to advance their technology initiatives.
The findings were presented by Dean Lacheca, VP Analyst at Gartner, during the IT Symposium/Xpo held in Gold Coast, Australia, continuing a series of Gartner trend briefings that in the past have covered a dozen technologies set to reshape the industry. The firm groups the trends around three interrelated shifts: the adoption of AI agents, security that needs to adapt continuously, and a reorganization of how governments deliver their technology services.
This isn’t a prediction that every government will adopt these technologies at the same pace. Gartner frames the trends as areas public-sector leaders should evaluate when deciding on investments and operating models.
AI Agents Are Starting to Raise a Governance Problem
The first trend Gartner identifies is what it calls the “domino effect” of agentic AI.
Agents differ from a conventional chatbot because they’re designed to pursue goals and use tools or external systems to complete different steps of a task. Depending on how they’re configured, they can look up information, execute operations, or coordinate several actions with varying degrees of autonomy.
In a government agency, they could be used to help with internal tasks, software development, document management, or certain citizen-facing services.
Gartner believes governments are moving from experimenting with artificial intelligence toward larger-scale projects, and that agents could become part of that process.
The trickiest issue comes up when a system stops limiting itself to drafting or summarizing information and starts stepping into administrative processes.
An incorrect response from an assistant can be reviewed by a person before it’s used. An agent connected to corporate applications may be able to carry out several operations before that review even happens.
That’s why the firm insists rollout must come paired with governance, employee readiness, metrics, and process redesign.
| Trend | Change it brings for governments |
|---|---|
| Agentic AI | Moving from isolated tests to governed systems built into processes |
| Cybersecurity | Replacing periodic reviews with more continuous adaptation |
| Post-quantum cryptography | Preparing systems and data for a future cryptographic transition |
| Public-sector workforce | Improving AI literacy and digital skills |
| Technology services | Reviewing operating models and responsibilities |
Gartner recommends developing outcome-driven AI plans and setting metrics to check whether investments are actually delivering value.
That distinction matters in the public sector. Adding an AI model doesn’t by itself prove that a procedure has become faster, cheaper, or more reliable.
Questions also arise around traceability, human oversight, data protection, security, and accountability for decisions. Their importance grows when AI takes part in services that directly affect citizens.
Cybersecurity Stops Working as a Periodic Review
The second trend involves a situation that’s less visible to citizens but essential to keeping public services running.
Gartner believes governments need to move from security based mainly on periodic reviews and compliance to a strategy able to adapt continuously to changing threats.
Artificial intelligence shows up here too.
Attackers can use it to speed up certain tasks, while defensive teams rely on automation and AI-assisted analysis to detect and respond to incidents.
The firm also cites two other factors: geopolitical tensions and the future arrival of quantum computers capable of affecting some of the cryptographic systems currently in use.
That doesn’t mean quantum computing can break the encryption used by governments today across the board.
The challenge is preparing far enough in advance to replace vulnerable algorithms when needed, especially in systems that store information that must stay protected for many years.
That’s why Gartner includes crypto-agility and preparation for post-quantum cryptography among its recommendations — the same seven-question checklist security vendors like Redtrust are urging organizations to work through before migrating.
In practical terms, crypto-agility means avoiding a situation where an organization gets stuck with cryptographic algorithms that are hard to locate or replace. Before migrating, it’s necessary to know which applications, devices, certificates, and protocols use each technology.
The transition won’t simply be a matter of installing an update.
Large government agencies can accumulate in-house applications, third-party software, legacy devices, certificates, public-key infrastructure, and industrial or specialized systems over decades. Identifying all of those dependencies can turn into a considerable undertaking.
Gartner also recommends automating security operations where possible, improving visibility into the digital supply chain, and strengthening risk management around vendors and data sovereignty.
Technology Modernization Also Depends on Employees
The third trend focuses less on specific products and more on how public technology departments are organized.
Gartner argues that the arrival of new technologies and demand for different skills will force a rethink of how IT services are delivered within governments.
Training stands out as one of the priorities.
In a Gartner survey conducted in January 2026 among 1,219 CIOs and IT leaders, 78% of respondents from government organizations said they expected to invest in their employees to advance their main technology initiatives.
The firm recommends prioritizing AI literacy, modernizing workforce skills, and greater organizational flexibility.
This part may draw less attention than autonomous agents, but it will shape how they’re used.
An employee using AI needs to understand not just how to write prompts for a model. They also need to know when to check a response, what information can be entered, which tasks shouldn’t be delegated, and how to spot results that sound convincing but are wrong.
Technology leaders will need additional skills to evaluate vendors, set permissions, control access to data, and decide which operations an agent can carry out automatically.
AI adoption thus brings new responsibilities on top of automating some existing tasks.
From Public Chatbots to Agents Connected to Administrative Systems
The three trends Gartner points to end up converging.
A relatively simple public chatbot can answer questions using previously published information. A more advanced agent could query different systems, pull up information tied to a case file, and help complete a procedure.
That second scenario offers greater possibilities, but it also raises the stakes of an error or improper access.
Governments would need to determine who authorizes each action, what information the agent can access, how its activity gets logged, and at what point a person must step in.
Security can’t be bolted on at the end of the project, either. If agents have access to corporate tools, their credentials and permissions become part of the attack surface that needs protecting.
That’s why Gartner brings artificial intelligence, cybersecurity, and workforce organization together under a single view.
The shift the firm describes for 2026 and beyond isn’t simply about adding more AI to public services. The question will be deciding where it can be used, how much autonomy it should get, and what controls it needs once it starts acting on real systems.
For governments, the difficulty may lie less in accessing AI models than in adapting legacy applications, data, processes, security, and employees to use them in a controlled way.
That work will also determine how far agents can go. A government with fragmented information, poorly defined permissions, and hard-to-integrate systems will have a harder time automating processes than one that knows its data, maintains well-managed interfaces, and can precisely determine who’s authorized to do what.
That’s why Gartner’s trends point to a different phase of public-sector digitization. AI keeps gaining ground, but its adoption is starting to shift attention away from technology demos and toward less visible issues: governance, identity, security, cryptography, integration, and training the people who will have to work with these systems.
via: Gartner
