利用我们的敏捷和DevOps炒作周期来指导技术投资

Gartner Research How PlatformEngineering TeamsCan AugmentDevOps With AI Manjunath Bhat, Cameron Haight, Bill Blosen How Platform Engineering Teams Can AugmentDevOps With AI 8 January 2024 - ID G00801906 - 14 min readBy Analyst(s): Manjunath Bhat, Cameron Haight, Bill BlosenInitiatives:Software Engineering Practices; Build a World-Class Software EngineeringOrganization AI is poised to shape the next frontier of DevOps. Softwareengineering leaders driving platform engineering initiatives shouldaugment DevOps with predictive and generative AI to improvedeveloper experience, enhance software delivery workflows andoptimize the software delivery infrastructure. Additional Perspectives SummaryTranslation:HowPlatformEngineeringTeamsCanAugmentDevOpsWithAI(13 March2024)■ Overview Key Findings Organizations are already using AI coding assistants, AI-augmented testing toolsand AIOps platforms to optimize specific activities within DevOps. However, reducingoverall lead time requires identifying and overcoming constraints across all phasesin software delivery.■ Generative AI presents new opportunities to reduce developer friction and improvedeveloper experience in multiple phases of the software development life cycle(SDLC). Examples of friction include inadequate understanding of codebase andtime spent in debugging, code reviews and root cause analysis.■ Improving the efficiency of software delivery workflows requires optimizing andharmonizing all stages of the SDLC. Examples of inefficiencies include long buildtimes, analyzing build pipeline errors, change impact analysis and slow incidentresponse.■ AI offers advantages beyond traditional automation to help product teams managetheir software delivery infrastructure in a reliable, sustainable and cost-effectivemanner.■ Recommendations Software engineering leaders driving platform engineering initiatives should: Identify AI use cases for systemic improvements in the SDLC by discovering andprioritizing constraints in software delivery workflows. Watch out for “AI washing”from vendors, and avoid overengineering with AI techniques when traditionalautomation options suffice.■Improve developer experience by supporting AI-augmented workflows that reducecognitive load, and help developers to achieve flow state across development,delivery and operations phases.■Enhance the efficiency of software delivery by augmenting DevOps workflows withAI-enabled optimizations to shorten the feedback loop for each activity in the SDLC.■Optimize software delivery infrastructure by providing self-service access to AI-enabled infrastructure management capabilities as part of internal developerplatforms.■ Strategic Planning Assumption By 2027, the number of platform engineering teams using AI to augment every phase ofthe SDLC will have increased from 5% to 40%. Introduction Software development has been one of the leading use cases for generative AI with thegrowing adoption of AI coding assistants and the launch of ChatGPT. In the 2023 GartnerIT Leader Poll on Generative AI for Software Engineering, 52% of IT leaders said theyexpect their teams to use generative AI in software development.1A survey of Gartnerpeer community members also reveals that 61% of software engineering leaders areexcited about generative AI’s potential for code generation.2 However, just because developers write code, it does not mean they spend most of theirtime writing code. Developers on average spend anywhere between 10% and 25% of theirtime writing code.3,4,5The rest of the time goes into reading specifications, writingdocumentation, doing code reviews, attending meetings, helping co-workers, debuggingpreexisting code, collaborating with other teams, provisioning environments,troubleshooting production incidents and learning technical and business concepts — toname just a few. Therefore, focusing on writing code and ignoring the rest of the DevOpsvalue stream can expose other inefficiencies in the development life cycle rather thanimproving overall performance. As a result, our clients are starting to take a broader view of AI-enabled use cases inDevOps and asking questions like“What are the opportunities for AI to shapeDevOps/DevSecOps within a three-year horizon?”and“How can we use AI across the SDLCas part of agile and DevOps workflows?.” The rapid evolution of centralized platforms and integration ofAI/ML in every stage of the software development life cycle —from ideation and planning to production deploymentmanagement — will revolutionize software engineering. — Sandhya Sridharan, Head of Engineers’ Platform and Experience,JPMorgan Chase Platform engineering teams will play a critical role in answering these questions becausetheir remit includes helping development teams to improve delivery speed, softwarequality and developer experience at scale. They should understand the needs of theiractive and potential platform consumers to get a unique perspective into the challengesfaced by multi