SOAS 2025: Misalignment Means Mediocrity
In the rush to modernize, many organizations are deploying AI to optimize outdated processes—without questioning whether those processes still make sense at all. It’s a classic case of automating a fax machine instead of just using email.
As we move through 2025, the pressure to deliver faster, more secure, and more efficient applications is intense. Yet complexity, legacy practices, and misalignment keep undermining AI’s full potential. To stay competitive, organizations must align strategies across security, automation, and deployment. Without that alignment, the risk of stagnation grows.
That’s the theme that arose from our analysis of responses from 659 technology decision makers across the globe. We asked about AI, about automation, about delivery and security, and what concerns, frustrations, and blockers are slowing their organizations down.
The results are not surprising, as our first reaction to a game-changing technology like AI is to use it to improve existing environments.
That’s why it’s important to consider that such a disruptive technology is an opportunity to rethink those environments—from operations to security, from application delivery to microservices—and ask ourselves if the way we’re doing things is really the best way to keep doing things, albeit much faster.
Let’s take a look at some of those “things” right quick, shall we?
Security focus: Prioritizing AI over microservices
Security tends to follow the buzz—AI is the new shiny object, so it naturally receives the most attention. But the forgotten corners of your architecture can be just as vulnerable, if not more so.
Data insight: AI deployments often receive top-tier security due to their direct interaction with sensitive data, while microservices—vital to modern apps—are overlooked.
Implication: Shielding AI is crucial but leaving microservices under-protected opens the door to breaches. These services often handle user authentication, data exchange, and core logic; a weak link here can threaten the entire ecosystem.
Manual operations amid automation ambitions
We love to talk about how AI and automation will transform everything, yet many teams still rely on manual steps. It’s like dreaming of driverless cars but insisting on push-starting them first.
Data insight: Although AI promises automated tasks like traffic optimization, nearly 29% of teams are still mired in script-writing, and 56% rely on human operators to kick off processes.
Implication: Legacy methods and manual interventions choke the pipeline. Even the most advanced AI can’t deliver results if the infrastructure relies on time-consuming, error-prone manual steps.
Deployment delays: Stuck in outdated processes
In a world that glorifies “continuous everything,” it’s ironic how many teams still use processes that require multiple manual approvals and ticketing. This not only delays deployments but also saps team morale.
Data insight: Traditional deployment practices, such as reliance on human operators and cumbersome ticketing systems, cause significant delays—23% cite ticketing integration as a primary automation roadblock.
Implication: Modern CI/CD pipelines are built for speed, but outdated processes slow them to a crawl. If you’re pushing for continuous delivery yet still depending on manual approvals, you’re missing the point of agile deployment.
Complexity: The silent barrier to application delivery automation
We tend to blame slow progress on a single big problem—like a security gap or an old database. In reality, it’s the combined weight of countless small complexities that can grind your automation efforts to a halt.
Data insight: The challenges in automating application delivery—across APIs, tasks, languages, and time zones—are widespread and evenly distributed, revealing systemic complexity.
Implication: Fragmentation across multiple tiers and services adds layers of manual effort and risk. Streamlining APIs, reducing redundant tasks, and unifying technologies can drastically reduce complexity, unlocking genuine automation benefits.
Final thoughts
Misalignment is the real enemy—whether it’s security mismatches, outdated manual steps, or legacy deployment holdovers. AI can optimize old processes, but it can’t fix fundamental misalignment on its own. The path forward requires not just plugging in new tools but questioning if the underlying workflows are still relevant. Embrace modernization with a clear strategy, or risk automating yesterday’s problems.
You can dig deeper in this year’s research by clicking this helpful, handy link and getting your own copy of the complete State of Application Strategy 2025 report.