How Showpad slashed code review time 65% with CodeRabbit
2.3x faster onboarding (44 days down to 19)
46% reduction in cycle time
54% acceptance rate on critical issues flagged
65% decrease in human review workload
Showpad is an AI-native revenue effectiveness platform that bridges the gap between sales, marketing, and buyers to help revenue teams close deals. With approximately 200 engineers working across diverse repositories, including legacy codebases and modern microservices, platform reliability is critical. Bugs don't just cause minor inconveniences. They impact revenue for customers who rely on the platform to present and close business.
The company’s leaders on AI adoption, Jeroen Minnaert, VP of Software Architecture, and Geoffrey Beertens, Principal Engineer, knew they needed to modernize their workflow to accommodate a larger team and diverse codebases after their recent merger with Bigtincan.
Previously, the team had tried to solve these problems by forcing GitHub Copilot top-down across the organization. "We pushed Copilot on everyone and then nobody used it," Jeroen recalled. They learned that simply buying a tool and telling developers to use it doesn't work. Adoption has to be driven by the engineers' perceived value.
Next, they took a different approach, a grassroots initiative. They allowed developers to choose their own tools through an initiative where engineers could share "war stories" and trial AI agents that actually solved their daily problems. The result was transformative: a 65% reduction in human review workload, with cycle times reduced from 44 days to 19 days. When the trial period ended, engineers flooded leadership with requests to buy the tool and one in particular rose to the top. It was clear this wasn't a top-down mandate; it was something developers had chosen for themselves.
As part of this grassroots initiative, many devs embraced CodeRabbit, an AI-powered code review tool that engineers actually wanted. After seeing substantial improvements in review quality and speed during the trial, the Showpad team adopted it widely. This allowed developers to focus on architecture and business logic when reviewing code, rather than catching linting errors and minor issues.
Before adopting CodeRabbit, Showpad’s engineering team identified a critical bottleneck: the code review process. With around 150 engineers actively committing on GitLab, relying solely on human reviewers was becoming unsustainable. Engineers were spending valuable time context-switching to review peers' code, leading to reviewer fatigue.
With dozens of merge requests flowing through the system weekly, engineers faced constant interruptions. The pressure on the team was growing, as every review meant breaking the flow state to check code, often resulting in superficial reviews that threatened to miss critical issues or delay shipping features.
All said, working with both modern and legacy applications post-merger created a steep learning curve. Reviewers often had to evaluate code in unfamiliar repositories. New engineers struggled to get up to speed quickly, and understanding the full impact of changes in legacy systems was difficult without deep institutional knowledge.
Solution: Developers picked their own tools, including CodeRabbit
The team evaluated tools across four key areas: coding, review, testing, and documentation. Ultimately, they found they didn't need a separate tool for every category. They settled on just two: a coding agent and CodeRabbit for the code review workflow.
Why CodeRabbit stood out
It outperformed other AI review tools by leveraging more context
CodeRabbit offered a balanced blend of AI-powered comments and traditional static analysis. It provided actionable insights without overwhelming the team with noise. "We were quite happy with the mix of functionality coming from CodeRabbit," Geoffrey shared.
Over time, CodeRabbit’s feedback became increasingly consistent and reliable. Developers especially valued CodeRabbit’s precise, context-aware suggestions, the kind that catch subtle but important oversights. CodeRabbit would surface practical issues, such as missing fields. “Those kinds of comments are most appreciated,” Geoffrey explained. The feedback felt relevant, actionable, and trustworthy.
Developers fought to keep CodeRabbit
The ultimate validation came when the CodeRabbit trial period ended. "There was a period in time where the trial ran out and we hadn't settled the commercials. We got a lot of feedback from engineers, urging us to adopt the product immediately." Jeroen said. The grassroots approach had worked. CodeRabbit was what the developers wanted.
Why Showpad loves CodeRabbit
Instant summaries and diagrams that accelerate onboarding
CodeRabbit generates summaries and sequence diagrams for most merge and pull requests, giving new team members instant context. This feature proved invaluable for legacy codebases with sparse documentation. "CodeRabbit really helps guide engineers who are new to the code, as it will add a diagram, which helps them explain what's being implemented," Jeroen said.
More feedback, less context switching
The data showed a surprising trend. The human workload dropped significantly. "The amount of comments and suggestions that we're seeing has actually increased," Jeroen noted. He said engineers are more able to stay in their flow state with CodeRabbit. "The pressure on your peers has gone down dramatically because you're not context-switching all the time," Jeroen shared.
54% acceptance rate = High-quality, actionable feedback
The team tracked acceptance rates to ensure the CodeRabbit wasn't just generating noise. They found a 54% acceptance rate on critical issues and over 45% on major and minor issues. The team also tracks post-deployment defects as a quality benchmark, ensuring CodeRabbit’s adoption improves speed without sacrificing reliability.
Results: Game-changing time savings
65% decrease in human review workload
By letting CodeRabbit handle the first pass, engineers were freed to focus on high-leverage work. "The real value of something like CodeRabbit is that you have a shorter cycle time between committing and pushing code and then getting code review remarks from your reviewer," Jeroen explained. “The review workload of human beings has gone down by 65%.”
46% reduction in cycle time
With faster reviews and less back-and-forth, Showpad achieved a 46% reduction in cycle time. This velocity is critical for their post-merger integration and maintaining a competitive edge.
37% increase in review suggestions
More feedback led to issues being caught earlier. Developers began learning from CodeRabbit's suggestions, improving code quality before a human ever reviewed it.
2.3x faster onboarding (44 to 19 days)
The combination of their coding agent and CodeRabbit made a tangible impact on the team’s growth. New engineers became significantly more productive, thanks to CodeRabbit, which helped them understand the codebase and catch mistakes in real time.
With CodeRabbit, Jeroen said, “our engineers can focus on the decisions that actually move the business forward.”
CodeRabbit = A new standard for engineering excellence
Before CodeRabbit
- •Reviewers stretched thin, burnt out from constant context-switching.
- • Post-merger code spanned modern microservices and unfamiliar legacy repos.
- •New engineers took 44 days to ramp, and a prior top-down AI rollout had failed.
After CodeRabbit
65% reduction in human review workload.
- 46% reduction in cycle time; onboarding cut from 44 to 19 days.
- 37% more review suggestions surfaced, with a 54% acceptance rate on critical issues.
For Showpad, the results speak for themselves: shorter cycles, happier developers, and a platform that scales as fast as their ambition. By empowering their engineers to choose the tools they trust, Showpad has ensured that their AI strategy isn't just a corporate initiative. It’s a competitive advantage.
Showpad didn't just buy another tool. CodeRabbit helped them raise the bar on their culture of continuous improvement.
Ghent, Belgium and Chicago, USA
https://www.showpad.comLanguages
JavaScript, TypeScript, PHP, Go
Challenge
A post-merger mix of modern and legacy repos, and a previous tool that never stuck were slowing Showpad's review process down.
