The numbers at Freshworks reflect an engineering organization that stopped operating on assumptions and started improving on evidence.
Throughput across the organization increased by 40%, a significant productivity gain achieved while maintaining quality. In some business units, the boost in features released reached 20 to 25%. Rework and maintenance, one of the most costly and demoralizing drains on any engineering team, fell by up to 50%. Cycle time improved by at least 10%. Code review time dropped 8 to 10%, as teams gained better visibility into pull requests, feedback cycles, and
AI-assisted code review insights contributing directly to how fast work moves through the SDLC.
One data insight in particular changed behavior across the organization. Hivel showed that review time dropped when the code being pushed was smaller. That single finding drove a deliberate push to keep
pull requests tight. Over two quarters, around 55% of all PRs stayed under 50 lines of code. Work moved faster, not because the team grew, but because they could finally see what was slowing them down.
The cultural shift was equally significant. "Data-driven reviews are now a natural part of our cadence," Swaminathan said. "And Hivel is the common source we rely on for those discussions." Teams started celebrating improvements because progress was visible and objective. When something slipped, Hivel helped leadership understand the reason fast, without guesswork.
What's Next for Freshworks
The most important number Hivel surfaced at Freshworks wasn't a cycle time metric. It was proof of what AI was actually contributing to production.
"One thing Hivel made clear is that suggestions accepted in Cursor don't tell the full story of AI adoption," Swaminathan said. "We are actively measuring AI-authored code versus manually written code and what that means for our investments. Hivel helps quantify that by correlating AI code with speed and quality."
"Today, every CFO and CEO wants clear ROI on AI investments," Swaminathan said. "Hivel gives us that in a board-ready way. That's critical to any organization navigating the AI transformation."
Freshworks has a clear path through 2026 and 2027: maximize engineering potential, keep engineers focused on quality work and innovation, and get more to market faster than competitors. The ability to measure
what AI is actually doing to engineering performance, verified and board-ready, is how they intend to stay ahead.
RESULTS SNAPSHOT
Freshworks improved Time-to-Market by 10% across the organization. Rework and maintenance dropped by up to 50%. Features released increased 20 to 25% in key business units. Code review time fell 8 to 10%.
.svg)
.png)
