Workleap uses DX for the AI SDLC: Removing bottlenecks and proving ROI
Workleap achieves 4x throughput and automates 25% of PRs using AI agents, with zero quality loss
Overview
Workleap, an HR tech company, quadrupled engineering velocity across 120 developers by using DX telemetry to measure and defend AI investments. Workleap runs on a 15-year-old codebase and had been experiencing constraints and rising token costs. Engineering leadership used DX to pinpoint review bottlenecks, automate routine checks, and achieve 4x weekly PR throughput.
“Without the data from DX, executive discussions around our $3 million token spend would have been purely defensive. Having real-time data allowed us to prove that moving faster was delivering tangible value to the business.” Guillaume Roy, Chief Product and Technology Officer, Workleap
About Workleap
Workleap is a Montreal-based software company founded in 2006 and bootstrapped to more than $100M in ARR. It builds two world-renowned products. Workleap is the AI agent that helps managers build engaged & high-performing teams. ShareGate helps IT teams get their Microsoft 365 ready for AI by understanding what’s in their tenant, cleaning it up, and keeping it under control. The company serves more than 15,000 organizations worldwide with a team of 350 employees.
About DX
DX is the leading engineering intelligence platform, trusted by high-performing tech organizations to measure and improve developer productivity, developer experience, and the business impact of AI investments. Developed by leading researchers, DX combines qualitative developer sentiment with quantitative SDLC telemetry to help companies baseline performance, remove friction, and quantify ROI.
Navigating legacy limits and review friction in a 15-year-old stack
For nearly two decades, Workleap had no formal engineering telemetry and relied solely on high-level Jira insights that provided limited visibility. As engineering leadership sought to modernize the software development lifecycle across two primary product divisions, Workleap Platform and ShareGate, existing operational constraints created delivery bottlenecks:
- Zero visibility: Workleap operated for 15 years without formal engineering performance tracking, which made it difficult for leadership to understand delivery friction and systemic bottlenecks.
- Code review bottlenecks: Mandatory dual-human reviews created delivery delays, slowing cycle times on routine pull requests and capping developer throughput.
- Architectural drag: Their legacy desktop software was built on older frameworks, which limited tooling adoption and slowed AI integration across team workflows.
- Increasing token spend: Finance was scrutinizing the engineering team’s growing AI token spend, but the team lacked data to prove its ROI.
To break through these constraints, leadership set an aggressive target: double PR throughput within 12 months with the same team size, while maintaining or improving code quality. Guillaume Roy, Chief Product and Technology Officer at Workleap, explains, "If we continued to go at this speed, we would simply fall behind in the next 10 years. We needed to rethink the whole product organization.”
Instrumenting the process for AI-assisted engineering
Workleap selected DX to kickstart its transformation toward that 12-month goal. By leveraging DX’s real-time visibility into pull request cycle times, engineering leaders were able to pinpoint code reviews as the primary bottleneck. These insights empowered Workleap to begin automating its process. The team set up pull request classifiers for low-risk code, cut out redundant human reviews, and let developers assign routine tasks directly to AI agents. “We used DX to instrument velocity and quality before we touched anything else, which meant we could see the real bottleneck instead of guessing at it,” Roy said. “20 years in software and this is the first time I’ve run an engineering org on this kind of data. I won’t go back."
The impact was immediate. By October, five months ahead of schedule, Workleap achieved its 12-month productivity goal. ShareGate met its 2x velocity target, reaching 7.1 weekly PRs per engineer. Meanwhile, Workleap Platform reached 4x throughput, consistently delivering 12.7 to 15.3 PRs per engineer per week.
Measuring and defending AI spend with DX
As AI usage accelerated across the engineering organization, tracking the relationship between AI spend and developer throughput became critical. To maintain cost efficiency during rapid adoption, Workleap introduced its own “elasticity metric” that connects DX to its infrastructure data, measuring rising AI token expenses directly against throughput gains. Under this model, teams that increased their velocity were required to balance their increased token costs accordingly.
Workleap’s AI initiative started with a $1M token budget, which quickly grew to a projected $3M+. Engineering leadership had to explain that gap to executive stakeholders, and for the first time it had the data to show the extra spend was delivering favorable productivity returns. Roy shared, “Without the data from DX, executive discussions around our $3 million token spend would have been purely defensive. Having real-time data allowed us to prove that moving faster was delivering tangible value to the business.”
Workleap continues to scale DX across its product development lifecycle, expanding telemetry into product management, design validation, and feature adoption workflows to ensure upstream discovery keeps pace with accelerated engineering output. Roy explains, “Shipping code is no longer our bottleneck, it now outpaces what product management and design can define, validate, and measure. We are using DX telemetry to unlock that next frontier.”
Published November 16, 2026