Artificial intelligence has officially crossed the threshold from experimental tool to business essential. A new survey by Zogby Analytics for Prove AI reveals that most companies have already embedded custom AI systems into their operations—but while the ambition is clear, so are the roadblocks.
Administrator
According to the findings, 68% of organizations have AI solutions in active production, and 81% are investing at least $1 million annually. Nearly a quarter are spending over $10 million, signaling that AI is no longer just hype—it’s a full-scale transformation.
Leadership is evolving to reflect this shift. More than 86% of companies now have a dedicated AI leader, often titled Chief AI Officer. These executives are nearly as influential as CEOs in driving strategy, with decision-making authority split almost evenly between the two roles.
Data Still the Biggest AI Roadblock
Despite the rapid uptake, the journey to scalable AI isn't smooth. Over half of business leaders say training and tuning models has been harder than expected. The most persistent pain point? Data. Quality issues, copyright concerns, limited availability, and inconsistent validation are slowing progress and undermining results.
Around 70% of companies report delays in at least one AI project, with data problems cited as the leading cause. It’s a stark contrast to the optimism around AI governance, where nearly 90% of leaders claim strong oversight. That gap between perceived control and operational reality suggests many are overestimating their readiness.
AI Applications Evolve Beyond Customer Service
While virtual assistants and chatbots remain popular (55% adoption), businesses are now exploring more technical and operational uses. Software development is the top use case at 54%, followed closely by predictive analytics and fraud detection (52%). Interestingly, marketing—once a launchpad for AI adoption—is falling out of favor.
Generative AI is also taking center stage. 57% of organizations now prioritize it, though most are combining generative models with traditional machine learning to get the best of both worlds.
Among the tools of choice, OpenAI’s GPT-4 and Google’s Gemini lead the pack, with DeepSeek, Claude, and Llama also gaining traction. Most companies rely on multiple large language models, embracing a multi-model approach to balance strengths and avoid vendor lock-in.
A Shift Toward Hybrid and On-Prem AI Infrastructure
AI’s infrastructure story is also changing. While 88% still rely on the cloud, there's a notable shift toward on-premises and hybrid deployments. Two-thirds of business leaders say non-cloud environments offer better efficiency and security, and 67% plan to move training data back in-house. At the heart of this trend is data sovereignty, which 83% of respondents list as their top deployment priority.
The move toward internal infrastructure underscores a growing desire for control—over data, compliance, and operational reliability.
