CloudBees has announced the acquisition of Launchable to boost GenAI and DevSecOps. The deal also sees Kohsuke Kawaguchi, founder of the Jenkins project, and Harpreet Singh return to CloudBees. Neither side has disclosed the amount paid.
Anuj Kapur, President and CEO of CloudBees, “Bringing CloudBees and Launchable together represents two milestones for the field of software development.
“First it brings the first AI-augmented Test Intelligence capability to any DevSecOps platform. Second, it marks the return of a true open source visionary to CloudBees to continue the mission-oriented efforts around software-based digital transformation in Enterprises.”
Why Launchable?
This is easy to answer on two fronts, acquisition of products and the return of former CloudBees staff to the company.
Returning staff
This deal sees Kohsuke Kawaguchi rejoin CloudBees. As the founder of the Jenkins Project, Kawaguchi is hugely influential in the DevOps community. When he left CloudBees to create Launchable four years ago, it raised questions as to why he couldn’t have created the company as a project within CloudBees.
It also sees Harpreet Singh’s return to CloudBees. His experience in product design at CloudBees, Atlassian, and, latterly, Launchable will be invaluable as CloudBees goes through its next product cycle.
Product gains
The key element here is Launchable’s technology. Testing has been one of the gaps in most CI/CD environments. While building pipelines to automate testing has happened, there hasn’t been a wholesale adoption of Continuous Testing (CT), despite the multi-year focus on DevSecOps. Rather than focus on the automation of testing, it is still a developer-led process. CloudBees says it takes up to 25% of a software developer’s time.
Launchable closes that gap. Unlike other test suites, it uses an AI co-pilot to create tests, evaluate the results, and speed up the test process. That use of AI extends to suggesting which tests should be used when code is committed and the ability to create dynamic tests. CloudBees believes that this will “slash dev-test iteration and accelerate ship times”.
CloudBees says that there are three immediate benefits to developers from this acquisition. They are:
Faster Dev-Test Iteration: A machine-learning driven predictive test selection reduces time spent on unnecessary tests and inefficient workflows. Teams know which tests matter, can automatically test upon code-changes, find failures faster and prioritize them, and find failing builds in a fraction of the time.
Efficient Test Triage and Analysis: AI-augmented testing helps teams effectively manage and analyze test failures, reducing the manual effort involved in identifying and fixing issues, thereby improving overall productivity.
Enhanced Visibility and Focus: AI-augmented testing improves visibility into test suite performance and test health, ensuring that teams can focus their efforts on fixing critical issues quickly and efficiently.
Developers want AI support for testing
The question of whether to rely on AI for writing and delivering code has been quite a topic over the last year, especially among low-code/no-code vendors. The main concern is the accuracy of the code and how that code is arrived at.
For example, the common claim is that using GenAI to write code means that the code and prompt are stored in and used by the GenAI engine. That makes the code and any flaws available to others. The problem is that this is all anecdotal, with no firm evidence that generated code has been used against an organisation.
The second concern is the quality of the code. This is much more provable. The concern is about the quality of the code being generated and whether it can be trusted.
When Enterprise Times asked Michael Beckley, CTO and co-founder of Appian, about AI-generated code, he replied, “We’ve discontinued most of those experiments right now because the models simply aren’t good enough. Just because the code might be syntactically correct, it may not be conceptually correct. And so we have not seen the kinds of efficiency gains that we were hoping for from writing code with AI.”
But what about testing?
In Stack Overflow’s 2024 Developer Survey, conducted in July, 46% of developers said they are interested in using AI to test code. When it comes to adoption, 81% of developers expect to see AI integrated into their workflow for testing.
Launchable has some numbers from its own customers, such as BMW and GoCardless. Both reportedly say they saw gains in time and quantifiable cost savings. According to the company, “they reported a 50% reduction in machine hours, a 90% reduction in test execution times and a 40% reduction in build times.”
Those are significant numbers. What is missing is a wider qualitative study to look at the benefits but also the potential risks. What are the costs associated with integrating AI? How do processes need to change? Which use cases see the most gains, and which are marginal? Most importantly, what lessons have customers learned, both good and bad, especially around changing culture to adopt AI?
Enterprise Times: What does this mean?
Acquisitions are made for various reasons, such as access to new technology, speeding up R&D, acquiring key talent, access to new customers, and even eliminating competition. In this case, there is no obvious downside to CloudBees acquiring Launchable.
This will give CloudBees customers even more capabilities within the product. It is also likely to increase the productivity gains they already get from their DevSecOps processes.
This reinforces CloudBees’ and its customers’ move to DevSecOps and will help deliver more secure and trusted code faster. With all the pressure on development teams, this has to be a major bonus.
