Case Studies / Dapta
How Dapta, a Voice AI Company, 10×'d Their PR Velocity Using Optibot
When almost everyone on Dapta's team started pushing code, PR volume jumped 8–10×. Dapta replaced slow manual reviews and an unreliable AI tool with Optibot — getting deep, full-context reviews in 1–2 minutes on every PR while Optimal AI Insights tracked their velocity in real time.
PR Velocity Increase
8–10×
As AI let engineers and non-engineers alike ship code, Dapta's PR volume jumped 8–10× — and Optibot reviewed all of it.
Time to Full Review
1–2 min
Every PR gets a deep, full-context review in one to two minutes, replacing slow manual reviews and an old, unreliable AI tool.
PRs Reviewed in 90 Days
2,700+
Across 16 active seats — roughly 170 PRs per engineer — every line is checked, verified, and scanned for security issues and tech debt.
PR Velocity Increase
8–10×
PR velocity increase as the whole org started shipping.
Time to Full Review
1–2 min
Deep, full-context review on every PR.
PRs Reviewed in 90 Days
2,700+
PRs reviewed in 90 days; every line checked.
See why engineering leaders at high-growth companies use Optimal AI
"Having Optibot has been like having a dedicated engineer that is just doing code reviews. Every line of code we ship, Optibot checks it, verifies it, and finds security issues. It became a critical part of our workflow almost immediately."
Felipe Gomez
Chief Product & Technology Officer, Dapta
Dapta is moving fast. Its platform deploys voice agents that let SMBs put AI to work across sales and operations. The way Dapta builds has fundamentally changed. The development cycle has compressed so much that there's no longer a separate designer, project manager, or product manager handing work down the line. Engineers, product owners, product specialists, and even product leadership prototyping in Node and Python all ship code directly. That pushed Dapta's PR volume up 8–10×.
That velocity created a review problem. With so many contributions a day, code and design started to drift — "everything looks different," as the team put it. Manual reviews couldn't keep up, and a code-review AI tool Dapta had been using fell short on quality. After adopting Optibot, every line of code shipped is now checked, verified, and scanned for security issues and tech debt in one to two minutes per PR.
Paired with Optimal AI Insights, Dapta now sees how its engineering velocity is changing in real time, streamed directly from GitHub.
Everyone (and Their Agents) Started Shipping Code, and Reviews Couldn't Keep Up
As AI unlocked shipping for the entire org, PR volume exploded and the review layer fell behind:
- Everyone ships now — with AI, the dev cycle collapsed: engineers, product owners, and specialists all write and merge code, pushing PR volume up 8–10×
- Design and code drift — with dozens of contributions a day across a large multi-repo codebase, consistency suffered. Some contributors were strong on design, others weren't, and the design system wasn't being enforced uniformly
- Reviews couldn't scale — manual code reviews were too slow to ship to SMB customers fast, and a prior AI code-review tool fell short on the depth and quality Dapta expected
"What started out as just the engineers using Optibot for code review has now moved into product owners, product specialists, and pretty much anyone that ships code at Dapta."
Felipe Gomez
Chief Product & Technology Officer, Dapta
Without a fast, trustworthy review layer, Dapta couldn't safely convert its new shipping velocity into shipped product.
Reviews in Under 2 Minutes, and Velocity Tracked Across 16 Engineers
Dapta integrated Optibot as the automated reviewer on every PR, starting with its most active, most complex repo and extending across the codebase. Alongside it, Insights turned GitHub activity into a live view of engineering productivity.
Reviews in 1–2 minutes
Optibot delivers deep, full-context reviews in one to two minutes per PR. The low latency and overall speed removed the review bottleneck entirely.
Every line checked
Optibot checks every line of code shipped — verifying it, finding security issues, and consistently surfacing tech debt and maintainability concerns.
Enforced standards across repos
Dapta codifies its design-system and best-practice standards once and has them enforced automatically on every PR, keeping a fast-moving, multi-contributor codebase consistent without manual policing.
Insights for real-time velocity
Streaming GitHub data shows how velocity is changing team-wide, per engineer, and including agentic productivity. Leadership uses it daily, including in 1:1s, and the team can ask Optibot about engineering productivity directly inside Insights.
"We started out just using Optibot inside GitHub, but then eventually we used all of the plugins as well as the skills available within Optibot in the marketplaces. Now every line of code shipped, Optibot checks it, verifies it, finds security issues, and consistently finds tech debt."
Felipe Gomez
Chief Product & Technology Officer, Dapta
8–10× More PRs Shipped, Every One Reviewed in Minutes, Quality Intact
Optibot became a permanent part of how Dapta builds, and it spread far beyond the engineering team:
- A dedicated reviewer on every PR — Optibot became core to the workflow almost immediately, acting "like a dedicated engineer that is just doing code reviews"
- Reviews in 1–2 minutes — removing the latency bottleneck and letting Dapta ship to SMB customers fast, even at 8–10× PR volume
- From a focused trial to org-wide — what began as a trial on the most active repos expanded to 16 of 17 seats active org-wide (94% utilization), now used by engineers, product owners, product specialists, and anyone who ships code
- Real-time productivity visibility — Insights is part of leadership's daily routine, including 1:1s, giving Dapta a live view of team-wide and agentic productivity, rework, PR size, and velocity per engineer
- Deep engagement — the team regularly gives Optimal AI product feedback because they want to use Optibot even more than they already do
"I'm seeing the Insights data every day now. It's been very useful in our one-on-ones — to compare and let people see their own performance."
Felipe Gomez
Chief Product & Technology Officer, Dapta
The Impact in Numbers
Before and after metrics for Dapta's team using Optimal AI
Real numbers verified by the leaders using the tech.
Metric
Before Optibot
After Optibot
Improvement
PR Velocity
Manual reviews couldn't keep up as everyone began shipping
Every PR reviewed regardless of volume
Review Turnaround
Slow manual reviews; prior AI tool lagged
Deep, full-context review on every PR
Review Coverage
Inconsistent; quality gaps with prior tool
Every line checked, verified & security-scanned
Code & Design Consistency
Drift across a fast-moving, multi-contributor codebase
Standards enforced automatically on every PR
Adoption
Engineers only
Engineers + product owners + specialists
Productivity Visibility
No real-time view of velocity
Insights streams GitHub data — velocity per engineer & agentic productivity
By the numbers — last 90 days (Mar 17 → Jun 17, 2026): 2,704 reviews · 2,634 PR summaries · 143 comment replies · 16 of 17 seats active (94.1% utilization). All activity from the Dapta-Tech organization.
Cut cycle time by 50% and get visibility into engineering productivity
Start reviewing PRs faster, catching issues earlier, and shipping with confidence.
