ChatGPT's public release in 2022 kicked off an ongoing conversation about generative AI and the future of work. Since then, as enterprise generative AI adoption has skyrocketed, exactly how it's adding value is still being debated. To better understand how companies are leveraging generative AI and perceive its value, Altman Solon, in collaboration with Cowen, surveyed over 500 IT buyers across companies in North America, Europe, and APAC. The findings provide a nuanced look at generative AI adoption, an evolving story that reflects both growing maturity and recalibrated expectations around return-on-investment (ROI).
We observed large-scale generative AI adoption across business functions, especially in software development and back-office functions. Nearly 80% of respondents report tangible benefits linked to generative AI use. However, most of these gains are realized in productivity and innovation, not yet in profit growth or cost savings, indicating that the technology works. Still, its scale of impact and pricing may need recalibration. Data security continues to pose challenges for many organizations, showing that adoption growth is still tempered by caution. The findings suggest that generative AI delivers measurable value and is emerging as a transformative enterprise technology, though the extent of its commercial impact is still taking shape.
However, the pace of the rollout has been slower than expected. In 2024, 83% of respondents predicted that they would be using generative AI by 2025, a target that has yet to be reached. This eight-percentage-point gap underscores the operational complexities of scaling generative AI within large organizations. This tension between high adoption intent and slower execution reflects a market still transitioning from experimentation to broader implementation. As adoption broadens and implementation matures, enterprises are also rethinking why they use these tools.
From cost savings to innovation
The motivations for adopting generative AI products are changing. In 2024, 68% of respondents said their primary goal was cost reduction; in 2025, that figure fell to 50%. Meanwhile, motivations tied to innovation rose from 55% to 64%, suggesting strategic advantages associated with generative AI tooling. "Expediting processes" remains the top driver, cited by 87% of respondents. This shift away from cost-cutting toward innovation highlights how generative AI tools are increasingly associated with high-value, strategic outcomes.
While early adoption focused on automating existing tasks to save time and money, companies today are experimenting with agentic AI. These systems can autonomously execute complex workflows to reimagine how work gets done. Our findings show that 25% of enterprises already use agentic AI and 37% plan to adopt it within the next three years. By enabling AI agents to act across workflows rather than within isolated tasks, enterprises are beginning to see how automation can translate into measurable outcomes: faster go-to-market timelines, more agile operations, and innovation at scale. Together, these findings suggest a pivot from near-term efficiency to long-term capability building.
Concerns around data security and unclear value remain
Despite widespread adoption and more advanced use cases, data security remains the top barrier to implementation, cited by 71% of respondents. Other common concerns include unclear value realization (38%) and model accuracy (34%).
Encouragingly, concerns over “tool history”, the perceived immaturity of AI tools, have eased slightly, decreasing by three points from 2024. This suggests increasing trust as solutions mature and enterprises develop AI expertise. Despite these concerns, adoption is expanding, reflecting a more deliberate, value-driven phase of enterprise experimentation.
Enterprise adoption is broad-based, but value realization is uneven
Adoption has increased across nearly every business function, with notable gains in customer service, software development, and back-office functions such as marketing, sales, legal, and HR. These areas, where repetitive tasks and large data sets dominate, have proven ripe for automation and augmentation.
Enterprises are tracking where generative AI is creating value. The most common performance metrics are full-time employee (FTE) time saved (63%), cost savings (62%), and customer issue resolution time (46%), all pointing to tangible efficiency gains. However, while 78% of respondents report measurable improvements in at least one operating KPI, most of these benefits are concentrated in productivity and efficiency, not revenue or profit growth.
This pattern reinforces a trend seen throughout the survey: organizations are reaping operational gains from generative AI but have yet to achieve broad commercial returns. Our findings show 91% of respondents report productivity and efficiency gains, while only 45% report commercial impact. As adoption broadens and capabilities mature, the next challenge for enterprises will be linking these productivity gains to measurable business outcomes.
The path forward: from experimentation to execution
As the enterprise AI landscape matures, success will depend less on early adoption and more on effective integration, measurement, and governance. The next wave of leaders will be those who treat AI not as a cost-saving experiment but as a strategic lever for innovation, differentiation, and growth.
Download the key insights from our report to learn more.
