AI hype hits a wall when the data doesn’t deliver
Companies are pouring money into AI for IT operations, but most projects are still far from maturity. A global survey of 1,200 business leaders, IT leaders, and technical specialists found that while spending and confidence are rising, only 12% of AI initiatives have been fully deployed.
The report, authored by Riverbed, suggests that optimism at the executive level is colliding with challenges in data quality, tool complexity, and everyday IT performance.
Most AI projects remain in early stages, with only 12% fully deployed across the organization.
Big budgets, slow rollouts
Organizations nearly doubled their AI investments over the past year. Many leaders and technical specialists agree the spending is delivering returns, with most saying outcomes have met or exceeded expectations. Even so, most projects remain stuck in pilot or development phases. Executives often express confidence in their strategies, while technical teams highlight ongoing issues with data, integration, and scaling.
This split helps explain why progress feels uneven. Budgets are growing, but systems are not yet ready to run AI at scale.
Tool consolidation accelerates
The survey found that organizations use an average of 13 observability tools from nine vendors, creating fragmented environments. Leaders acknowledge this patchwork adds overlap and slows teams when problems arise. Almost all respondents are consolidating, and 93% say they are willing to switch vendors to make it happen. The main drivers are integration, productivity, and alignment with business strategy.
That level of openness suggests the vendor landscape could look very different in the next few years, as companies move toward platforms that cover more ground with less complexity.
Unified communications under pressure
Hybrid work has made unified communications tools central to daily operations. Employees spend a large share of their week on video calls, chat, and collaboration platforms. Despite that reliance, performance remains a frequent complaint.
Users continue to report dropped calls, connectivity problems, and inconsistent quality. These issues show up in help desk queues, where UC tickets take longer than average to resolve. Responsibility for UC systems is often divided across multiple teams, which makes it harder to identify problems or apply consistent fixes.
OpenTelemetry gains traction
OpenTelemetry, the open-source observability framework, is moving toward becoming a standard. A large majority of companies report at least partial adoption, and most expect vendor support to be mandatory within two years.
Business leaders are more likely to call it a strategic priority, pointing to its value in improving visibility and preparing for automation.
Networks carry the weight
The ability to move and share AI data is another pressure point. Respondents named cost, security, and performance as their top concerns when shifting workloads across cloud, edge, and on-premises environments. While public cloud and edge storage are expected to grow modestly by 2028, use of on-premises data centers is likely to decline.