The rise of practical AI: The search for business value
According to the most recent Aberdeen AI research, only 4% of businesses are not using AI in some capacity. Whether through deliberate deployment or simply by using software where AI is embedded in everyday business tools, the technology has moved from pilots and trials to the core of daily operations.
From our report: “As AI reshapes IT operations, organizations are finding practical ways to integrate it into daily workflows and infrastructure. From September to October 2025, Aberdeen Strategy & Research conducted a 15-minute online survey among 281 IT decision makers who are personally involved with tech purchase decisions related to AI for their org. The study examined how IT professionals across North America, EMEA, APAC, and India are using AI today and the outcomes they are achieving.“
However, this ubiquity comes with potential hazards. There’s high likelihood of an AI bubble driven by circular investments and a frenzy of vendors pushing unready, immature AI capabilities onto users who neither want nor need them.
As AI hype continues to accelerate, businesses must choose between irrational AI or practical AI. Understanding the distinction between these two approaches is the defining factor between successful AI driven modernization and expensive stagnation.
Practical vs. Irrational AI
The difference between a successful AI initiative and a failed project doesn’t solely lie in the technology itself, but rather in the approach a business takes. In our research, we tend to see businesses either follow AI hype cycles irrationally (what we call Irrational AI) or take a more measured approach (as in practical AI). Let’s define these further.
Irrational AI is characterized by a “just do it” mentality. It’s fueled by the fear of missing out rather than business logic. Organizations following this approach like to talk about “transforming everything” while lacking a coherent plan to transform anything specifically.
Irrational AI also tends to rely heavily on public large language models (LLMs), exposing themselves to data privacy risks and generic outputs. In many ways, these businesses seem to view AI as “magical”—a black box that will solve unspecified problems simply by existing. This approach is the primary fuel for the looming AI bubble, creating noise without signal and expense without return.
Practical AI, conversely, is defined by a clear focus on using AI to achieve specific, tangible business goals. Rather than relying completely on generic public models, practical AI leverages internal data to build AI capabilities that are focused on what matters for a business. It is also deployed as part of a well-planned, phased rollout rather than a “just do it” implementation. Most importantly, practical AI is governed by defined success metrics where a business can actually show how and where AI is working for them.
The Drivers of Adoption
Aberdeen research has shown consistent growth in dedicated deployments of AI (meaning not just capabilities embedded in existing software solutions). And when we look at the pressures pushing these businesses to adopt AI, we see that critical and “practical” concerns are the top drivers.
The primary driver, cited by 52% of organizations, is the need to combat increasing cybersecurity threats. As attackers utilize automated tools to breach defenses, businesses are turning to AI to analyze patterns, detect anomalies, and respond to threats at machine speed.
The second major driver is rising cloud costs (34%). While AI itself requires compute power, it is increasingly being used to optimize infrastructure, identify waste, and streamline cloud resource allocation.
Finally, lack of application development efficiency (25%) is pushing companies toward AI. In an agile world, the ability to write, test, and deploy code faster is vital for success. AI-driven coding assistants and automated testing frameworks are becoming essential for maintaining development velocity.
Strategies for Readiness and Infrastructure Hurdles
Recognizing the need for AI is one thing; being ready to deploy it is another. The top strategies for enhancing AI readiness include the use of dedicated AI platforms (35%) to develop, train, and deploy models, ensuring a structured environment for innovation.
With lack of AI expertise a constant challenge, 34% of businesses are prioritizing in-house AI skill development and hiring initiatives for AI/ML talent. Furthermore, recognizing that AI is only as good as the information it’s fed, 25% of organizations are focusing on data management enhancements.
However, our research also uncovered a number of key challenges businesses face on the road to successful AI. The ability to integrate AI with existing legacy IT platforms is a significant challenge, as this is vital for improved usability and engagement.
Also, the high cloud capacity required to run advanced models, along with the associated costs, poses a barrier to entry. Database capabilities also serve as a challenge; traditional databases are often ill-equipped to handle the vector search and high-speed retrieval needs of modern AI applications.
The IT Engine: Support, Services, and Benefits
Nowhere is the impact of practical AI more visible than in the realm of IT support and services. This sector has become the proving ground for practical use cases, including faster issue resolution, improved analytics, and the automation of repetitive workflows.
Looking at the benefits realized in IT service management (ITSM), we find that 50% of organizations utilizing AI in ITSM report an improvement in IT productivity. By offloading level-one support tickets to conversational AI and automating routing, IT staff are freed to focus on complex, high-value strategic initiatives. This trickles down to the wider organization, resulting in improvement in end-user productivity (reported by 35% of respondents) and an increase in end-user satisfaction (33%).
For businesses that navigate the infrastructure hurdles and adopt a practical mindset, the impact of AI on broader business metrics is often positive. Looking at businesses that have adopted AI for a year or longer, we find positive impacts on system uptime and reliability, incident response time, and overall user satisfaction. The top benefits seen by adopting businesses include improved process efficiency, better automation of tasks, and increased performance.
Analyzing businesses with longer adoption times, the research shows a clear maturity curve. These organizations experience better AI accuracy, likely due to the fine-tuning of models on internal data over time.
They also report improved employee and customer engagement with AI—moving past the initial friction of adoption to a state of collaboration. And these businesses report improved IT results in critical areas like performance, uptime, and return on investment (ROI).
By focusing on internal data, defining specific metrics, and targeting high-impact areas like cybersecurity and IT productivity, businesses can insulate themselves from the hype. The future belongs to the practical—those who use AI not because it is trendy, but because it solves a problem, saves a dollar, or secures a future.
