26-06-2025
Sorry, DevOps: Garbage Data Can Only Generate Garbage AI Outcomes
Savinay Berry is the Executive Vice President and Chief Product Officer for OpenText.
As artificial intelligence (AI) continues to evolve, we've seen more and more of its capabilities and limitations. We've witnessed AI perform tasks once deemed futuristic, reminiscent of scenes from sci-fi pop culture, such as Hanna-Barbera's iconic 1960s cartoon, The Jetsons. In this animated series, nearly every aspect of life is automated, from housework and meal prep to fashion, depicting a world where AI plays a central role in daily activities.
While exaggerated, the show prompted me to contemplate exactly what role AI can—and should—play in helping product leaders deliver the next generation of innovation. Specifically, I began to wonder: How can AI be integrated into DevOps strategies as more than just a tool, but as a strategic ally—an extension of software development teams.
The benefits of AI in DevOps—from automating testing and deployment to improving resource management and enhancing security—are driving increased investment. The "2024 Developer Survey" from Stack Overflow revealed that 76% of developers are using or planning to use AI tools, up from 70% last year. Notably, 81% cited productivity gains as the biggest benefit, while 62% valued accelerated skill development.
As an innovator, I've already seen AI refine operations, simulating human intelligence across workflows. AI and automation offer unparalleled opportunities to reimagine software delivery. It streamlines resource-intensive areas such as testing, code-writing and deployment, all while ensuring compliance and security.
AI and automation can also autogenerate test scripts, adding ready-to-go test scenarios to plans in a single click, as well as creating anticipatory project reports to pinpoint potential risks that could jeopardize software quality. And this is only scratching the surface of what is possible.
So, What's The Issue?
The risks of not adopting AI are significant and go beyond the DevOps community—it is a larger CIO issue. According to a recent survey from OpenText, 96% of respondents are using, testing, or planning to explore AI across their organization. Moreover, 78% believe failing to leverage internal data effectively will squander AI's potential.
This seems promising, but the reality is more nuanced. In DevOps, the intersection of AI and operations goes beyond implementing advanced algorithms; it demands a robust foundation of organized, high-quality data. Without this foundation, achieving desired AI outcomes becomes a formidable challenge, and we risk stumbling at the first hurdle.
Cleaning Up Your Data Is More Than Fixing Spreadsheets
Understanding the role of information management means recognizing that AI thrives on high-quality data. The expression "garbage in, garbage out" applies here. If data isn't managed for accuracy and accessibility, desired AI results won't be achieved. High-quality data, on the other hand, sets the stage for success.
Consider a car engine: Removing deposits and sludge (inaccurate, outdated, irrelevant and incorrect information) reduces friction, while clean oil (large language models and AI) ensures smooth performance. Much like good car engine maintenance, effective information management ensures longevity and optimal performance, generating superior and enduring AI results.
Remember, even the most skilled data scientists and developers can't achieve optimal results without reliable data. And while many businesses understand AI's dependence on quality data, few feel ready to act on it.
It's a shame because only organizations that govern, unify and protect their information will unlock AI's full, transformative potential.
Optimize Your AI Initiatives With Quality Data
Is your AI engine running on premium data? For DevOps teams, maximizing value starts with a solid information management foundation.
Begin with these five steps:
1. Audit Your Data: Conduct a comprehensive review of all data assets (data logs, appdev dashboards, metadata, etc.) across cloud and on-premises storage. Identify data sources, formats and quality to build a knowledge base for AI workflows that can support faster application delivery, automated testing and intelligent code suggestions. Clean historical data also improves time-to-market predictions.
2. Set Data-Governance Standards: Adhering to data governance standards ensures that the applications they build comply with privacy regulations and industry standards. Establish clear data-governance protocols to safeguard privacy and ensure proper data management. Consistent, high-quality data flows are essential for reliable AI results, making effective governance paramount.
3. Implement Continuous Data Integration: Embrace an ongoing process to integrate disparate datasets into a unified format suitable for AI analysis. This continuous aggregation ensures AI assistants have a relevant and useful foundation for effective functionality. This can reduce the burden on not only application developers, but also quality assurance testers and managers, as accurate results can be assured every time.
4. Secure Data Flows: Prioritize data security and align your practices with regulatory requirements and industry standards. Implement proactive validation checks and AI-driven threat detection to maintain data integrity and security. Developers must incorporate security best practices, such as encryption and access controls, into their data handling processes.
5. Enhance Data Accessibility: Enable conversational search interfaces and AI assistants to access relevant datasets from multiple knowledge bases, when your information is clean. These tools can empower developers to optimize the software delivery cycle and reduce delivery times. With tools like this, clients I've worked with were able to maximize test coverage in less time and on (or under) budget, while enjoying greater access to high-level insights.
Elevating DevOps With AI Through Information Management
AI and automation have already shown their potential to enhance operations and productivity. However, the next step is sustainable and scalable AI. This evolution is essential for reimagining business information ecosystems and elevating people to become innovative leaders, rather than late adopters.
As technology leaders, we stand at the forefront of the AI revolution, and we must recognize that information management is not merely a support function, but the catalyst for AI excellence. Effective information management not only boosts AI capabilities, but it also encourages innovative, targeted and successful application development, delivery, execution and measurement.
Embracing information management as a cornerstone of AI-driven progress is the key to achieving the level of product excellence that once seemed like science fiction.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
