Businesses can accelerate DevOps maturity by harnessing the power of artificial intelligence (AI), specifically automation. However, for a company to truly reap the rewards of AI, it must move beyond tools such as Copilot and utilize more advanced solutions including TuringBots or AI agents. These will allow for stronger adherence to pipeline standards and greater productivity for engineering teams.
AI Agents: The Tool Empowering Tomorrow’s Engineering Teams
An AI agent is a series of prompts that leverage AI to carry out a role within the engineering team, such as the business owner or tester. AI agents act as assistants to help ensure the code adheres to all requisite security, quality and coding standards before merging with the main line and different environments, saving the developer precious time.
For example, if an engineer needs to verify their code, they can have an AI agent interpret, validate and make suggestions. Based on those suggestions, and with human approval, the AI agent can then build, rewrite or refactor whatever is missing, whether tests or quality gates. The faster feedback and reduced workloads possible with AI agents enable engineering teams to become proficient at churning out code, getting it tested, and moving it into different environments.
Likewise, as these AI agents ensure the code is high quality earlier, there will be fewer defects later in the software development lifecycle (SDLC), meaning less rework. This will, therefore, massively improve the productivity of the engineering teams, translating into faster time to market. Ultimately, this AI-assisted DevOps environment will empower teams to achieve a higher level of DevOps maturity at an unprecedented pace.
How Businesses Can Reorganize Teams in Light of AI
Building an AI-first engineering team will require legacy teams to undergo top-to-bottom change. Five years ago, a head of DevOps ran all the tooling to support how code moves, testing, etc. There is now an ongoing trend in the DevOps space for this position/title to transform into one focused on platform engineering.
These new platform engineering leaders are accountable for enabling a variety of AI-related objectives in their organizations. They own all the platforms that engineers use daily. These leaders must also account for AI enablement, whether overseeing all AI models the business has integrated with or managing the costs of these models through insights into token usage (tokens are how the different LLMs charge their clients).
Like the responsibilities of the platform engineering leader, a dedicated team must be accountable for caring for and feeding the platforms and tools used for DevOps. This team must own and support the AI agents because as applications mature, the AI agents need to change. Likewise, they will be responsible for encouraging AI adoption across engineering teams.
Even the simplest tools like Copilot, which has a low entry bar, require significant change management activity. Nevertheless, teaching senior engineers new ways of working can be difficult. Businesses should lean into education and reward mechanisms, such as organization chain management initiatives, to help reskill existing teams. Despite these best efforts, getting engineers to adopt new AI tools won’t happen overnight, it could take three months or longer.
Furthermore, companies must prepare for the inevitable restructuring of their engineering teams. Businesses will likely need to reduce headcount once AI agents begin automating and streamlining more processes in the SDLC. Currently, most engineering teams consist of a product owner, a scrum master, a businessperson, several developers and a few testers. Soon, these teams will only need a product owner, one engineer (such as an engineering lead or architect), a tester, and a business person. At the same time, enterprises do not need highly skilled engineers adept at building anything, rather their priority is someone with more domain expertise and understanding of the business.
Intentionally Implementing AI
Despite the undeniable productivity and cost-efficiency benefits of AI, many companies are still hesitant and even fearful of this technology and the implications it may have on their engineering teams. While AI will cause disruption, this is not uncharacteristic of all technological innovations at some point or another. Still, organizations must not immediately jump in feet first. They must find ways to incorporate AI by experimenting and testing. Moreover, businesses can enlist the help of experienced partners to craft incremental technology and business strategies.
