F5 recently released its annual State of Application Strategy (SOAS) Report, highlighting a shift in the use of artificial intelligence (AI). The report indicates that AI has moved from an experimental phase into production workloads. The findings emphasise AI’s increasing role in organisational environments and the associated need for operational and security standards similar to those applied to other critical systems.
Approximately 78% of enterprises now independently manage AI inference, reflecting a preference for control over convenience. This trend is particularly relevant in the context of growing multicloud operations. Around 93% of organisations use multiple cloud environments, reflecting increased complexity in managing and delivering AI workloads securely.
AI’s use has expanded from experimentation to a core part of business operations. Organisations deploy an average of seven AI models, with a strong focus on inference. This shift has increased attention on governance of AI systems, with integration into application frameworks that require architectural and security considerations.
However, AI-as-a-service strategies are viewed by many organisations as carrying higher risk, with just 8% relying solely on public AI offerings. Most organisations instead use a mix of models, which requires routing and policy controls to manage cost, accuracy, and availability.
The hybrid multicloud approach is widely used for delivering AI workloads. In line with broader multicloud adoption, enterprises deploy applications across multiple environments, which requires governance across these systems. Integration across environments and consistent policy enforcement are used to support security and help reduce operational disruption.
As AI is deployed at production scale, security considerations have become more prominent. About 88% of organisations have encountered AI-related security issues, and most are preparing for the development of agentic AI systems. This has increased focus on governance mechanisms, particularly around prompts, tokens, and API-based control layers.
The report also notes an increasing focus on control at the prompt and token layers, prioritised by 29% and 23% of organisations respectively. Governance in these areas is associated with managing cost, performance, and safety in AI workload operations.
Overall, the State of Application Strategy Report provides data on changes in enterprise technology adoption. It highlights the continued development of AI use in production environments, the expansion of hybrid multicloud architectures, and evolving approaches to security and governance.
