It is rare in the world of government technology for budget forecasts to go up against macroeconomic gravity. However, 2026 promises to be an anomaly. Despite global pressure to cut spending and balance sheets, government IT departments – excluding the US – are bracing themselves for budget increases. The reason is clear, albeit risky: the boundless belief that artificial intelligence will modernise public services faster than they can grow old. However, behind the optimistic excel columns lies a serious challenge highlighted by Gartner analysts: the risk of getting carried away with novelty at the expense of technological fundamentals.
If we had to define the one key word that will open the wallets of government CIOs in the coming months or so, it would be ‘productivity’. It is the one that is becoming the new holy grail of digital government, pushing soft image metrics somewhat to the side. But the road to this productivity is a bumpy one, and the roadmap drawn by Gartner shows that decision-makers are faced with a dilemma: invest in trendy ‘Agent AI’ or patch holes in business processes?
The investment landscape: the Big Four and a new player
The figures from market analyses are clear. Governments are not planning minor adjustments, but a technological offensive. As many as 85% of IT directors in the public sector plan to increase spending on cyber security. This is understandable – in the age of hybrid warfare, a digital shield is a priority. Right behind security, however, comes a trio of innovations: artificial intelligence (80%), generative artificial intelligence (GenAI) (also 80%) and cloud platforms (76%).
It is particularly interesting to see how quickly the appetite for AI is evolving. Not so long ago, we were talking about simple chatbots. Today, 74% of offices have already implemented or plan to implement AI within a year, and the figure for GenAI is up to 78%. However, the real news that is electrifying the industry is so-called Agentic AI (Agent Artificial Intelligence).
Unlike the models we are familiar with, which ‘just’ generate text or images, AI agents have the ability to autonomously take action and perform complex tasks on behalf of the user. This is the promise of a civil servant that never sleeps. Not surprisingly, nearly half (49%) of government CIOs say they are interested in this technology. For the IT industry, this sends a signal: the government customer is looking for solutions that don’t just ‘support’, but actually ‘perform’.
Voice of reason: Between hype and craftsmanship
In this technological race, however, it is easy to become breathless. Arthur Mickoleit, chief analyst at Gartner, pours a bucket of cold water over the heated heads of enthusiasts in his predictions. While he acknowledges that agent-based AI is emerging as a powerful facilitator of transformation, he warns of a phenomenon we know all too well in the industry – the ‘hype cycle’.
There is a real risk that the media hype around autonomous agents will divert the attention of decision-makers away from technologies that are less media-savvy, but much more mature and necessary. We are talking about classic machine learning (Machine Learning) and business process automation (BPM).
For IT integrators and suppliers, there is a key lesson here: the role of the technology partner in the coming years will not just be to deliver the latest innovation. It will be to educate and bring the customer down to earth. Deploying a sophisticated AI agent on ‘dirty’ data or based on a non-optimised, analogue process is a simple recipe for a costly disaster. Innovation in the public sector must be an evolution, not a revolution detached from the foundations. Mickoleit rightly points out that it is this next, less glamorous wave of innovation that will be needed to meet the sector’s priorities.
‘Check’: The end of art for art’s sake
The change in narrative in government IT is also evident in the strategic objectives. The days when digitalisation was an end in itself are over. By 2026, technology is expected to ‘prove’ concrete results.
More than half of directors (51%) cite increasing employee productivity as their main goal. This is a drastic change in optics. Until now, it was common to talk about ‘citizen experience’ (CX) – this goal, while still important (37%), is giving way to the pressing need for internal optimisation. Governments are struggling with staff shortages and an ageing workforce. Technology is set to fill this hole.
In practice, this means that CIOs are now being asked for hard evidence of the impact of investments on the office’s mission. They have to demonstrate cost savings and real improvements in the quality of work. For software and service providers, this means they have to change the language of benefits – instead of ‘modernity’ and ‘innovation’, they have to talk about man-hours, automation of repetitive tasks and ROI. The pressure for results is high, especially as the public and budget overseers have years of investment in e-government in their minds and now expect dividends from these outlays.
Geopolitics enters the server room
The final, but perhaps most important piece of the puzzle for 2026, is digital sovereignty. IT purchasing decisions are no longer the domain of technical performance and price alone. Big politics have come into play.
As many as 55% of government executives expect changes in the way they interact with technology providers due to geopolitical tensions. The location of the supplier is becoming a critical factor, put on a par with the cost or scalability of the solution.
This is excellent news for local and regional technology companies. Nearly 40% of decision-makers plan to work more closely with suppliers based in their region. This is a clear shift away from an unreflective reliance on global hyperscales. Governments want to identify risks in their technology stacks, avoid reliance on a single supplier (vendor lock-in) and build resilience to future shocks.
Gartner experts emphasise that the identification of dependencies and risks is becoming the new standard. For the Polish and European IT sector, this opens a window of opportunity. The argument “we are from here, we are subject to the same law and we guarantee data sovereignty” is gaining as much weight as the technical specification of a server or the functionality of an algorithm.
What does the market need in 2026?
Looking at the projections for 2026, we see a picture of a sector that is ambitious but also cautious. Money is on the table, but it is money marked by high expectations.
The ideal IT project for government in the coming period is one that combines fire and water: harnessing the potential of modern AI (including generative and agent-based AI), but embedding it on a solid process foundation and a secure, preferably local or diversified infrastructure.
