Mention GenAI to most DevOps team members and they’ll tell you about their use of Code Copilots. While this was one of the first uses of AI in DevOps and is still the most widely practiced, AI has now been embedded deeply into all five phases of DevOps.
This article examines some of the lesser-known, but powerful uses of GenAI that will boost the productivity of the entire DevOps team.
Plan
As DevOps automation for build, test, and release improves, the planning phase is set to become much more important. Planning is not just about tickets, planning poker, resource allocation tables, and burndown charts, but it starts with the gathering of requirements at a high level. User stories are generally part of an epic and/or feature that represents a higher level of activity. An architecture plan is part of planning. So is a UX prototype and design brief; AI can help with all of these.
AI can review a User Story to ensure it’s complete and well-written. Think of it as Static Code Analysis for the requirements. AI can even help with creating acceptance criteria and dependency analysis, and break up a well-written epic or feature into individual User Stories.
All of these steps will improve the productivity of the Product Owner, make the communication of the intent much clearer and less ambiguous, and provide the foundation for the next few phases.
Build
In addition to code generation which is well documented elsewhere in the press, AI can assist both pro coders and low coders in other ways.
First, for pro-coders, AI can generate unit tests and help formulate an SOQL query based on natural language interactions. They can get an explanation for code generated by the AI in addition to getting help troubleshooting issues with that code as well. Oh, and if you provide a corporate style guide in your workspace, the code generated by the AI can be made to conform.
There are also several goodies for low coders too. AI can generate formulas based on your needs and your org’s existing metadata. Even the most senior admins find it difficult to remember all the details about formulae and the other inner workings of Salesforce. Trailhead and Salesforce help are amazing resources, but just ask the build agent a question and the answer will be provided along with reference links to learn more or validate the answer.
All of these and more combine to accelerate development and ensure your solutions adhere to best practices for your company.
Test
Testing is a rich area of opportunity for AI. Gen AI can read a well-written user story and generate both manual and automated test scripts based on the intention of the story. This makes it possible to shift left and practice test-driven development. There should no longer be any doubt if a User Story is ready for integration.
If you’re like most Salesforce shops, you probably don’t have the test coverage you aspire to. This is typically due to a lack of resources. Instead, most Salesforce teams do manual testing. Good news for those shops, an exploratory testing tool can watch a subject matter expert conduct a manual test and turn it into an automated test. AI can then start with the happy path test and enhance it to include negative testing and boundary tests.
And, let’s not forget User Acceptance Testing (UAT). Exploratory testing tools can help boost the productivity of your UAT sessions by making it simple for business users to provide feedback and exact steps to reproduce any defects.
Finally, if you have a library of tests written in one script language and want to convert them to another, just ask a Test Agent. Why would you want to do that? Maybe you’re tired of your Selenium tests breaking every time Salesforce does a release and you’d like to convert to a more robust, modern scripting language like the Robot Framework.
Release
When you think of Release Automation, you likely think about using orchestration engines, but GenAI can help here as well. The Release Agent can help build automation scripts and even generate release notes from the User Stories that actually made it into the release.
Operate
Instead of calling it “Monitor”, we prefer to call the final stage of DevOps “Operate” since there’s much more to operating a Salesforce org after the release than simple monitoring. How about training users, for example? The Operate Agent can generate training documentation and act as an expert on using the Application. Simply create a workspace with all your past and present release documentation and the Agent can answer any questions your users might have on how to use the new features.
Summary
Copado provides AI Agents that deliver on all of these benefits, from Plan through to Operate.
There are two major points we hope you take away from this article.
- GenAI can boost the productivity of every member of your DevOps Team and even help your business users. Second, many of the benefits of AI are based on well-written, complete User Stories, Features, and Epics.
- Gone are the days when your development team was sitting in the same small room and able to ask the Product Owner for clarification. Proper documentation in the planning phase not only improves the efficiency of the team but also provides valuable documentation for the life of your applications.
So, don’t fail to plan. It’s easy with AI tools that are specifically designed for each and every phase of the DevOps process.