A new industry report shows that while enterprises are advancing from AI experimentation to full-scale production, measurable returns on investment (ROI) remain limited for most organisations. According to the 2025 REVelate report from Domino Data Lab, 88% of enterprises have improved their ability to deploy AI at scale, but nearly 60% expect less than 50% ROI from machine learning and generative AI (Gen AI) initiatives.
The report surveyed over 300 C-level executives, VPs and directors across North America and Europe.
“It’s clear that the AI hype cycle has peaked. Enterprises are done experimenting, and they’re racing to scale, but the payoff is still out of reach,” said Domino chief operating officer Thomas Robinson. “Our data validates what leaders know about and followers are learning: aligned leadership, flexible infrastructure, and integrated governance drive a straight line to AI business impact. These aren’t technical upgrades — they’re business imperatives that will define who sees real returns from AI.”
Organisations prioritise governance, tools, and cloud flexibility
The report indicates that business leaders are shifting focus toward building AI-ready infrastructure and establishing governance frameworks, instead of chasing short-term financial returns. Integrated AI/ML governance was ranked the most critical capability for executing enterprise AI strategies, cited by 66% of respondents. Tooling (49%) and hybrid cloud infrastructure (44%) followed closely.
Hybrid environments are becoming increasingly central to enterprise AI strategies, with 48% of respondents identifying hybrid cloud infrastructure as essential. Even among organisations with less formalised AI strategies, 42% prioritised hybrid approaches.
However, the pace of AI innovation is presenting new pressures. A majority of respondents (63%) reported difficulties keeping up with rapid advancements, while 59% cited unrealistic expectations from business stakeholders. Another 57% noted a lack of clearly defined AI leadership roles, underscoring structural challenges in managing AI growth.
High costs and sector differences underscore AI deployment challenges
The report also noted that cost continues to be a major barrier. Organisations face high expenses in acquiring AI talent (63%), addressing unexpected project costs (61%), and leveraging existing infrastructure (61%). The overall cost of developing and operating AI systems was a concern for 60% of respondents.
Sector-specific insights revealed varying levels of AI maturity and adoption. Life sciences companies showed the highest uptake of advanced AI technologies, with 98% using GenAI and 95% adopting agentic AI. Despite this, only 3% of these organisations expected more than 100% ROI from generative AI in the short term.
Financial services firms demonstrated the most developed AI governance frameworks, with 51% reporting formally appointed AI leadership roles. In the public sector, foundational challenges remain with only 23% of respondents in this sector had a defined policy for AI-related data access and use.
Recently, another survey found that many employees are hesitant to adopt new workplace technology, with one in seven refusing to use new tools altogether. Additionally, 39% of respondents identified themselves as reluctant users.
Read more: 51% of companies now use AI agents, finds new survey
