AI-powered employees gain momentum

Agentic AI heats up in 2025

Agentic AI – which promises semi-autonomous agents taking over whole processes – represents a leap forward in generative AI (GenAI) technology.

While at the start of 2025, this branch of GenAI is in its relative infancy, early movers are starting to develop new agentic AI concepts rapidly.

Deloitte predicts that 25% of enterprises using GenAI are expected to deploy AI agents in 2025, growing to 50% by 2027.

Dubbed ‘the new frontier’ of GenAI, PwC predicts agentic AI will contribute between $2.6 trillion and $4.4 trillion to the global economy by 2030.

Enterprises should investigate agentic AI, but leaders must be wary of industry hype. GenAI is still prone to misleading ‘hallucinations’ – where the outputs generated can be inaccurate. If agents apply these hallucinations to business decision-making or customer interactions, it could lead to negative consequences.

For example, AI news summaries hit the headlines for all the wrong reasons in January. The BBC complained to Apple about an AI feature that shared misleading headlines attributed to the BBC’s journalism. Apple was forced to suspend the AI news alert service as a result.

Having a ‘human in the loop’ for generative AI oversight is vital. Rather than having AI run freely on an overall task, consider how teams can use AI to carry out groundwork before a subject matter expert reviews the outputs.

But the pace of progression does not relent – and early adopters already have agentic AI trials underway. Some organizations are piloting AI agents to run credit checks, for instance. Crucially, humans ultimately oversee the end-to-end approval process.

Full-scale agentic AI adoption is still over the horizon, and there are more immediate factors for enterprises to face today.

However, the incoming EU AI Act will introduce a level of caution for many businesses this year. Some enterprises will look at AI through a different lens when they apply the criteria set out by the European Union.

For others, it will bring comfort, giving businesses clarity over the rules and potential issues while creating a clear legal boundary within which they can operate. It will also crystalize best practices around data governance and how AI tools should be used.

The desire from leaders to see a ‘return on investment’ (ROI) on AI is also a continuing theme for 2025, as businesses seek to understand the payback on their initial investments.

While 2024 saw many pilots and proofs of concepts developed, some AI projects are struggling to reach the next stage of developing robust business cases.

What is your organization’s biggest concern regarding AI adoption?

AI replacing human jobs
Ethical and bias concerns
Data security and compliance
Inaccurate AI-generated outputs (hallucinations)
No concerns,
we fully trust AI

Assessing AI challenges and blockers in 2025

Business leaders are broadly optimistic about AI – two-thirds (67%) believe it will fundamentally change the nature of their business in the next two years, according to KPMG.

However, nervousness around regulation and compliance requirements in AI will continue this year. The same study highlighted organizational data quality as the biggest anticipated challenge to AI strategies in 2025 (85%), followed by data privacy and cybersecurity (71%).

Enterprises without secure and clearly defined data sets will be concerned about creating potential litigation issues when putting AI tools in the hands of employees.

Human-centric application of AI puts people first

Despite being difficult to quantify, productivity is traditionally used as a success metric to determine the ROI of AI. But an over-reliance on this blunt metric in isolation masks the bigger picture.

Instead, consider additional measurements such as reducing employee stress levels or attracting and retaining new talent. These complementary numbers can add greater clarity to the ROI of AI, which in turn makes the case for further investment.

Measuring the value of AI with a more human-centric approach can also influence adoption.

Consider the wide range of capabilities and levels of understanding around AI across the workforce. A ‘one-size-fits-all’ ideology to AI adoption will not work as intended – and instead, it needs to be tailored to the individual, factoring in their levels of AI maturity and experience.

Remember that changing human behavior remains the biggest blocker to the AI tool adoption.

To create value for the business, leaders need to force a change in thinking – asking employees and team leaders to challenge how work can be done differently. For example – would using AI make a certain task easier to complete?

Nominating champions or AI evangelists is a powerful way to support the adoption of new tools. This creates a visible point of contact for anyone needing help, while acting as a driver for change.

What is the primary driver for AI adoption in your organization?

Automating repetitive tasks to improve efficiency
Enhancing employee productivity with AI-powered insights
Reducing operational costs through AI-driven processes
Improve customer experience
Reduce workloads and improve employee well-being

How Fujitsu enables AI-empowered employees

At Fujitsu, we take our customers through a co-design process to help them
understand the value they want to achieve with AI.

By helping our customers take a human-centric view, we support them in achieving their
long-term business goals and spot easy wins. Our consultative approach helps identify the
value of AI while taking a customer through the steps to implementation.

Fujitsu’s Kozuchi AI platform combines a wide set of AI technologies in one place to help
businesses make complex decisions. For example, it can be combined with video feeds in
a warehouse to minimize threats to safety, notifying supervisors with auto-generated alerts.