CloudBees Buys Launchable In GenAI, DevSecOps Play

This ‘separates us a little bit also from the competition,’ Shawn Ahmed, CloudBees’ chief product officer, tells CRN in an interview.

Cloud-native developer security operations platform provider CloudBees is boosting its quality assurance and generative artificial intelligence capabilities with the acquisition of Launchable – potentially bringing more QA work to its channel partners.

As part of the acquisition, returning to San Jose, Calif.-based DevSecOps vendor CloudBees are Kohsuke Kawaguchi and Harpreet Singh, Launchable co-founders and co-CEOs.

Kawaguchi is also known as the creator of the Jenkins open source automation server, and at CloudBees Singh helped to scale the Jenkins business to millions of dollars in annual recurring revenue (ARR).

Launchable “starts observing and learning the pattern behavior between the types of code changes developers are doing and the types of tests that will fail because of that code change,” Shawn Ahmed (pictured above), CloudBees’ chief product officer, told CRN in an interview. “Multiply that time saving with – from one developer to 10, 10 to 100, 100 to 1000, 1000 to 10,000 – and you can see how the multiplicative effect of those savings will have. … (This) separates us a little bit also from the competition.”

CloudBees Acquisition

CloudBees has a partner program for system integrators, resellers and other business models. The vendor has more than 500 enterprise users. The companies are not disclosing financial terms of the acquisition. The acquisition closed in July.

Code generation holds one of the biggest opportunities for partners in AI, Mainline President and CEO Jeff Dobbelaere told CRN in a recent interview.

The practice cuts across a variety of verticals, making it easy to scale that offering and meet the demands of mainframe customers modernizing their systems, Dobbelaere said. Mainline, which is not listed as a CloudBees partner on the vendor’s website, has been leveraging code-generation tools by IBM and IBM subsidiary Red Hat as part of this practice.

“We have a number of customers that have legacy code that they’re running and have been for 20, 30, 40 years and need to find a path to more modern systems,” he said. “We’re still in [GenAI’s] infancy, and the sky’s the limit. We’ll see where it can go and where it can take us.”

CloudBees’ Ahmed estimated that developers spend less than 30 percent of their time coding, mostly concentrating on activities including deployments, pipelines, tests and then observations. Launchable can take the time spent on coding up to 70 percent, with more deployments and commits in a day, he said.

The proliferation of AI assistants such as Microsoft Copilot and Amazon Q means more code volume and increases the importance of decreasing time needed for QA, he said.

“AI has been tremendously successful for coding scenarios, (but) this is one of the great use cases of applying smart AI to an area that is right within the workflow of developers,” Ahmed said. “That's how Launchable thinks about it and how CloudBees thinks about it as well.”

Sacha Labourey, CloudBees co-founder and chief strategy officer, told CRN in an interview that the additional QA capabilities should resonate with CloudBees partners because companies tend to use consultancies for that work.

“There is, obviously, a clear need to differentiate and then to show how you can do a lot more with any given cost to your customers,” said Labourey. “You get to solve much bigger problems. You get to tackle much more complexity.”

Services-led channel partners remain an important part of CloudBees’ go-to-market because they know how to solve customer problems and conduct digital transformation with new technology, he said.

CloudBees was already using Launchable internally, leading to a smooth integration process, he said. He said Launchable is especially useful for users leveraging large test suites and embedded computers, including with mobile phones and in the auto industry.

AI Here To Stay

Ahmed said that AI in code generation is here to stay, and that CloudBees developers have leveraged the emerging technology.

“It's going to get better very, very fast,” he said. “So I don't think we're going to see diminishing use of AI or machine learning. I think it's the opposite. We're going to see a very fast, high curve as quality, all of this, gets better.”

The issue for end users is figuring out where to apply AI for the best early return and most value. Now the market has matured more and users can ask better questions about leveraging retrieval-augmented generation (RAG) and adding proprietary data to improve models, he said.

As for the importance of open source in AI – a theme echoed by open source-focused vendors including Red Hat and Databricks – Ahmed said that “AI is only as smart as its training data, and so open source models, I think, are the answer for companies being able to really wire up their systems in a way that they can trust, the government will trust, and their customers will trust.

“That will continue to be a very important aspect of AI adoption moving forward,” he said.

Open source is also essential for experimenting with AI at scale and introducing AI as part of users’ infrastructure, Labourey said.

Improving The CloudBees Platform

With Launchable, the CloudBees platform should gain faster dev-test iteration, with a machine learning (ML)-driven predictive test selection meant to avoid unnecessary tests and inefficient workflows, according to CloudBees.

Users will have the ability to automatically test upon code changes and find failing builds in less time, according to CloudBees.

With Launchable, the CloudBees platform will gain greater efficiency in test triaging and analysis, according to CloudBees, founded in 2010. Users can better manage and analyze test failures with less manual effort in identifying and fixing issues.

And the combined platform should improve test suite performance and health visibility so that users can fix critical issues faster and more efficiently.

Singh and Kawaguchi co-founded Launchable in 2019. Singh’s resume includes about a year with Atlassian, leaving in 2019 as head of product and general manager of Bitbucket Cloud, according to his LinkedIn account. He worked at CloudBees for more than seven years, leaving in 2018 as vice president of product management and design.

Kawaguchi’s resume includes about 14 years with CloudBees, leaving the vendor in 2020 as chief scientist, according to his LinkedIn account.

CloudBees was a member of CRN’s 10 Hot DevOps Companies To Watch In 2022.