Most organizations now view agentic AI as essential to managing cloud environments, but few have figured out how to run it across the business. 

Recent research from global IT provider Unisys shows that 75% of organizations see agentic AI as essential to managing their growing number of cloud applications. However, only 23% have started scaling it across functions. The rest are experimenting or remain in the early stages of adoption. 

Unisys surveyed 1,000 senior IT and business decision-makers across eight markets in the United States, Europe and Asia-Pacific region.

The disconnect between expectation and execution hands channel partners a clear mandate: move customers from pilots to production. Doing so means guiding organizations through decisions about data, architecture, security, governance and cost, not adding another product.

“The enthusiasm around agentic AI is real because organizations recognize its potential to improve productivity, automate decision-making and help manage increasingly complex cloud environments,” Mike Thomson, CEO and president of Unisys, told Channel Dive.

Enterprisewide deployment requires confidence in the data, visibility across the technology environment and the right security and governance controls. Existing infrastructure also can get in the way. “Every company already has technical debt, some of it just deployed,” Thomson said.

AI workloads raise additional questions about cost, security, latency, redundancy and interoperability. Costs vary depending on the workloads involved, where they run and the infrastructure supporting them. 

“It’s one thing to build it, but it’s entirely different to orchestrate the production so it runs effectively and efficiently,” he said.

Vendors package more work for partners

Pilots can prove a use case. Scaling across a business is a heavier lift. And that’s where Omdia’s latest quarterly cybersecurity channel analysis offers context. The research firm — a Channel Dive sister company — said several cybersecurity vendors are formalizing partner-led consulting roles around AI security and platform adoption.

The vendors are aligning their go-to-market strategies around AI and using controlled access to advanced models to support consulting engagements spanning vulnerability discovery through remediation. Omdia said the aim is to accelerate platform adoption through services-led partners and reshape where channel companies provide value. 

The findings reinforce what Unisys identified from the customer side. Organizations often understand agentic AI’s potential, but many still need help choosing workable use cases, preparing environments and managing production deployments.

“This is where channel partners can provide significant value,” Thomson said. “They can help clients identify practical, high-impact use cases, establish the operational and security frameworks required for scale and integrate agentic AI into existing technology environments.”

For partners, the work can span assessments, implementation, integration, governance and ongoing operations. The productized consulting model Omdia described could also help partners build repeatable services across customer accounts instead of treating every deployment as a wholly custom project.

Vendors benefit as well. Partners can handle the integration and operational work needed to put more of a platform to work for customers, without vendors having to deliver every engagement themselves.

Omdia sees cloud marketplaces gaining ground as a key route to market for cybersecurity sales. The firm expects spending through hyperscaler marketplaces to grow 33.9% to $10.96 billion in 2026 and reach $30.99 billion by 2030. Partner-led marketplace transactions continued to expand during the first quarter, Omdia said.

The forecast measures cybersecurity technology spending, not consulting revenue. Still, that growth could give partners more opportunities to attach implementation and managed services to technology sales.

Confidence outpaces results

Agentic AI scaling challenges are tied to larger problems, according to Unisys. The foundations are in place, but the results lag.

Among respondents, 89% reported progress identifying priority use cases, 87% cited progress upskilling employees and 87% reported progress securing executive support. 

Yet, that preparation has not consistently translated into better performance. Nine in 10 organizations said they have the architecture needed to support large-scale, data-driven decision-making. Less than two-thirds said operational efficiency exceeds expectations. 

Talent shortages, cited by 27% of respondents, and data management and integration problems, cited by 26%, remain the top barriers. Meanwhile, just 42% of business leaders view cloud and IT as profit centers rather than cost centers, down from 82% in 2025.

Customers have not stopped investing. More than 9 in 10 respondendents in the Unisys study plan to maintain or increase agentic AI spending. The channel opportunity is to make those investments work in production and deliver results customers can measure.