包阅导读总结
1. 关键词:AI 实施、战略路线图、企业软件、效率提升、长期成功
2. 总结:本文论述企业应积极以战略方式实施 AI 以获取竞争优势,介绍了其在增强安全、提升效率方面的作用,强调构建坚实框架、以数据驱动、负责任推进及规划的重要性。
3. 主要内容:
– 企业正积极将生成式 AI 融入核心业务
– 2028 年 75%企业软件工程师将用 AI 代码助手
– 早期采用者有经验教训
– 构建强大 AI 策略的好处
– 增强安全态势
– 提升开发者工作效率
– 持续改进软件开发周期
– 确保长期成功
– 规划考虑复杂性和现有周期
– 以数据为中心修改软件周期
– 建立数据驱动的 AI 结构
– 用数据提升输出质量
– 协作利用数据仓库和湖
– 推进 AI 采用的步骤
– 建立 AI 指导委员会
– 制定负责使用的指南和流程
– 为多模型方法做准备
– 负责任地采用 AI
– 避免盲目实施的风险
– 促进跨团队战略讨论
– 注意成本管理
思维导图:
文章地址:https://thenewstack.io/creating-a-strategic-roadmap-for-effective-ai-implementation/
文章来源:thenewstack.io
作者:Brian Wald
发布时间:2024/7/18 17:15
语言:英文
总字数:1095字
预计阅读时间:5分钟
评分:88分
标签:人工智能整合,企业战略,开发安全运维,人工智能代码助手,数据驱动的人工智能
以下为原文内容
本内容来源于用户推荐转载,旨在分享知识与观点,如有侵权请联系删除 联系邮箱 media@ilingban.com
Enterprises are no longer experimenting with generative AI; they are integrating the technology into their core operations to drive efficiency, innovation and competitive advantage. Whether your company is an AI early adopter or just beginning its journey, your focus must be on integrating AI to deliver tangible results. Success requires a strategic approach that ensures the AI solutions you implement align with business goals and provide qualitative and quantitative benefits.
There’s no time to waste. By 2028, 75% of enterprise software engineers will use AI code assistants, according to Gartner, which defines them as tools that assist in generating and analyzing software code and configuration. That’s up from less than 10% in early 2023. Even though the adoption curve is steep, we are still early in the hype cycle. If your developers haven’t fully adopted an AI code assistant, they may be well-positioned to avoid some of the growing pains early adopters have experienced.
Building a robust AI strategy now can help you catch up to and even surpass your competitors in measurable ways. First, AI-powered tools can significantly enhance your security posture by automating threat detection and response. They can identify vulnerabilities and potential attacks much faster than traditional methods, helping to protect your systems and data more effectively.
AI is already proving its ability to enhance productivity and introduce new efficiencies across the entire developer workflow. Although code generation is a popular AI use case, GitLab research found that developers spend less than a quarter of their time writing code. Over 75% of a developer’s day-to-day work can be supported and made more efficient with AI through vulnerability explanations, code change summaries, automated tests and more. For example, writing tests is often time-consuming and lags behind the deployed code. AI can assist by generating tests in real time, ensuring thorough coverage and enhancing the overall quality of the code.
Additionally, a well-implemented AI framework that grants a holistic view of the software development life cycle can continuously improve itself as more data is collected and added to context windows. Enhancements in AI technology automatically integrate into your workflows, providing ongoing benefits without requiring significant manual updates.
Ensuring Long-Term Success
Implementing AI requires careful planning and consideration of the complexity involved and your current software development life cycle. Where will AI create the most efficiency? In testing and deployment, maintenance or another area?
A thoughtful analysis of your software development life cycle, including tools and processes, will provide the answers and help inform where to implement a robust AI system to deliver sustained value and how to iterate and expand over time. Creating a solid AI framework is crucial for long-term success. AI adoption requires modifying the entire software development life cycle to centralize data. Centralizing data allows it to be used across the organization to inform prompts and enables retrieval-augmented generation (RAG) techniques to enhance generative AI outputs.
Build a Data-Driven AI Structure
Data is the lifeblood of your organization’s AI efforts. Feeding it into your AI systems will allow you to supercharge the output quality for both your developers and your customers, and empower AI to help you measure success and identify areas for productivity optimization.
Additionally, as you integrate AI into your applications, prepopulating them with appropriate training data will ensure they are production-ready from Day 1. Now is the time for developers and data scientists to collaborate on using data warehouses and data lakes to facilitate access to training models and application usage.
Onboarding: Three Planning Considerations
The first step in your AI journey is creating an AI steering committee with a strong leader to review and recommend approaches specific to your organization’s needs. In some cases, this may be a chief AI officer. A CAIO doesn’t have to be an immediate addition to the C-suite; it may be a transitional title that a VP assumes to coordinate AI usage across teams. However, it should be someone empowered to look holistically at AI.
The AI steering committee should bring legal, security and engineering leaders together to build a structure for AI adoption. It’s important to remember that AI success isn’t possible without first addressing the privacy, security and legal requirements your organization may face and how AI adoption plays into continued compliance.
Next, guidelines and workflows should be established to ensure DevSecOps teams use AI responsibly. For example, by setting up automated testing, including using a security analyzer, to create a gating mechanism that ensures code is reviewed before it is promoted to production. And beware of shadow AI — the latest variation of shadow IT — where workers subscribe to their assistants while working on your codebase, leading to IP leakage.
Now is also the time to prepare for multimodel approaches. LLMs tuned for specific tasks are proliferating, making it possible to use multiple AI models across a range of applications. Enterprise systems are beginning to implement fine-grain controls to help DevSecOps pick the right LLM for the job — or automatically optimize the routing for a given task.
Approach AI Responsibly
With new power comes great responsibility. Companies that don’t use AI for software development risk falling behind their competitors that do.
However, leaders who rush into implementing AI without understanding its impact invite risks like security vulnerabilities, fines, customer attrition and reputational damage. To implement AI safely, leaders must gain visibility into their teams’ workflows to understand how AI can support them.
Organizations can ensure alignment between executive leadership and their developers on critical topics by fostering an environment where strategic discussions about AI are the norm, involving cross-team conversations with technical, legal and AI service providers. An AI transparency center is a resource that can help decision-makers navigate these conversations and issues with clarity and confidence.
Treat AI integration like cloud transformation — be mindful and deliberate. Avoid piecemeal solutions when possible, and focus on a comprehensive approach that ensures sustainable growth and improvement. Many AI vendors are moving from per-seat pricing to consumption-based pricing. Managing this line item carefully will be necessary, and an AI tool can likely help. Manage consumption-based spending and avoid toolchain sprawl by ensuring AI investments improve outcomes.
AI offers tremendous benefits, but success requires a holistic and strategic approach. The goal is to create an environment where your DevSecOps teams can thrive with AI superpowers built into their tooling, backed by the appropriate leadership structure to support their efforts and show business value.
YOUTUBE.COM/THENEWSTACK
Tech moves fast, don’t miss an episode. Subscribe to our YouTubechannel to stream all our podcasts, interviews, demos, and more.