Eight IT predictions for 2025 and how to prepare

AI. Cloud computing. Quantum cryptography. These technologies promise transformation, but how will these innovations impact the workplace in 2025? 

We assembled leading industry experts and Unisys technology leaders to explore what’s reshaping technology and business. Their insights are the foundation of our latest research report, “Top IT Insights for 2025: Navigating the Future of Technology and Business.”

This report provides practical guidance for your technology and business initiatives in 2025 and beyond. You can prepare for the road ahead by understanding these eight key findings.

#1: Natural language will become the AI-human interaction default 

As AI tools become a staple in the workforce, humans won’t need as much encouragement. Instead, AI will require more training to understand humans and become more intuitive.

At the same time, prompt engineering to create AI inputs will become obsolete as humans begin communicating with machines as comfortably as they do with other humans. AI will be woven into our daily lives in natural, intuitive ways. 

What you can do: Start transitioning to natural language interfaces in your AI applications and expand your AI deployment beyond web and desktop applications. 

#2: Small language models and edge computing are the next frontier

Privacy, security concerns and connectivity limitations prevent some data infrastructures and AI applications from connecting to large language (LLM) models. Combining LLMs and edge computing lets organizations process data closer to its source and maintain AI’s advanced capabilities for more secure, faster AI applications. 

“Give the technology an opportunity to be used more broadly in places where you don’t have to send the data to the cloud,” said a former health system CIO. “It can actually do the compute where it exists…. It’s going to go much more to the edge, and that’ll bring other players in.” 

What you can do: Consider the speed and security gains of edge computing, deploy small language models and combine these with edge for a secure, fast and efficient AI solution. 

#3: Hybrid cloud will ascend while cloud migration decisions will demand nuance 

The cloud growth rate is slowing as organizations question the cost savings of cloud migration. Depending on business constraints and needs, they’re reevaluating which workloads should be moved to the cloud, and which should stay on-premises.

Cloud providers are hiking prices to offset the expense of AI compute. Cloud costs like storage charges, premium services and data transfer fees are putting additional pressure on CIOs. A hybrid cloud approach can be the best of both worlds, with sensitive, latency-dependent or regulated resources kept on-premises and scalable, data-intensive and customer-facing applications moved to the cloud. 

What you can do: Implement monitoring of cloud usage and costs and weigh the economics of the cloud versus on-premises. 

#4: Post-quantum cryptography will move from theory to practice 

By 2030, quantum computing may give cyber attackers the means to break current encryption standards. Acting now is imperative because they could steal your encrypted data now and decrypt it once quantum technology advances. This “harvest now, decrypt later” tactic poses a significant data security risk. 

“We know that there is an immense amount of data out there that’s encrypted in a fashion that in the not-so-distant future is going to be worthless. So, we’ll have to change the encryption,” said a Columbia University cybersecurity leader. “The general business world has not focused enough attention on the risk of quantum.” 

In response to the quantum threat, the National Institute of Standards and Technology (NIST) has developed post-quantum cryptography standards “designed to withstand cyberattacks from a quantum computer.” As threats evolve, post-quantum cryptography will need to grow, too. 

What you can do: Use NIST’s post-quantum cryptography standards as a starting point and plan to update your standards regularly to protect data and prevent breaches. 

#5: Entry-level workers will be the model for an AI-enabled workforce 

AI can automate repetitive tasks, but that doesn’t mean they can replace humans. Entry-level workers who are digital natives are often eager to use AI tools. They can contribute valuable knowledge of how to introduce these tools effectively into workflows.

Human-AI collaboration is the future of work. AI will assist with repetitive tasks like data entry, analysis and reporting, allowing employees to focus on work requiring creativity and human judgment.

What you can do: Recognize that employees have different comfort levels with AI, so focus on it being a tool for humans rather than instead of them. Provide training so that all workers can benefit from AI tools. 

#6: A three-day onsite schedule will be the sweet spot 

Organizations want to optimize how employees work, whether in the office, remotely or through a hybrid approach. Hybrid work has emerged as the best of both worlds, balancing workplace flexibility and in-office connection. 

A hybrid sweet spot of three days a week in the office and two days remote is emerging as the most sustainable long-term solution. One benefit is appealing to younger workers, who value the in-person mentorship, networking and collaboration that occurs in the office. 

What you can do: Evaluate whether a three-day in-office hybrid approach is your organization’s sweet spot. Invest in group activities to foster collaboration and employee engagement and mental health programs to diminish feelings of isolation. 

#7: Energy-efficient AI will be a competitive advantage 

Massive data centers built for AI training and operations are straining the global energy grid. This has sparked growing discussion about AI’s environmental impact and the need for energy-efficient AI models. Organizations that develop sustainable AI practices gain a competitive advantage. 

“Practicing sustainable AI could be a differentiator,” said Unisys AI Portfolio Vice President Suzanne Taylor. “Look at the cost of the chips and training — only the big companies can train large language models. We’re already seeing AI optimizing processes in industries like cargo, leading to sustainability benefits. The future will involve green AI as a competitive edge.” 

What you can do: Establish sustainable AI practices, move some AI workloads to edge devices to relieve the data center burden and choose partners that prioritize energy-efficient AI solutions.

#8: Diverse global standards will shape AI compliance’s future 

As AI technology advances, regulatory frameworks are emerging to address critical concerns around data privacy, security, intellectual property and ethics. By 2026, key regions like the EU and US are expected to develop more comprehensive AI governance principles.

To abide by regulations, organizations must enforce strong AI governance policies and will need to tailor their AI models by region. The biggest concerns will be data privacy and security, particularly for LLMs and other AI systems that process sensitive data. This will make data sovereignty a priority and rigorous AI model testing and certification a likely mandate. 

“AI will have access to information, and the question will be how we will ethically manage and protect that data,” said a former Fortune 500 board member in the survey. “Many companies don’t have good data readiness. They don’t have a good data governance capability.” 

What you can do: Set clear guidelines now and train your employees so they can abide by regulations. 

Get more details on 2025 IT trends 

Staying ahead means understanding the technologies that will reshape business. Our report delivers executive insights and strategic recommendations to help your team navigate 2025's most critical IT developments. Grab the full report to discover three key factors that will drive success in the rapidly changing tech landscape. 

Download the report