GitLab 19.5 adds Recommend Reviewers Flow to pick the right merge request reviewer

Finding the reviewer is the first bottleneck

GitLab frames the problem as one of routing rather than headcount. According to the company, the expertise and capacity to review code usually exist inside a team, but nothing matches them to each merge request as it arrives. Authors are left to work out who knows a given part of the codebase, who has room in their queue and who is online.

GitLab says that leads to habits that slow reviews down:

• Assigning a whole group. Everyone is notified, nobody feels responsible, and the merge request waits for a volunteer.
• Defaulting to the same trusted reviewer. That person becomes a bottleneck and their own work stalls.
• Guessing across time zones. The chosen reviewer may have just signed off, so feedback arrives the next day.

Reviewers also pay a price, according to GitLab. Group assignments flood their notifications with merge requests they will never review, and the people everyone defaults to take on a growing share of the queue on top of their own work.

How Recommend Reviewers Flow works

The flow runs on GitLab Duo Agent Platform. A trigger starts it automatically when a merge request is created, whether the request is in draft or ready state, so suggestions are available from the start.

It then weighs each eligible candidate's workload, time zone, availability, and expertise as defined by approval rules and CODEOWNERS. The aim is to find the person best suited and best positioned to respond. The flow explains why it chose each reviewer, and suggestions appear in a dedicated panel on the merge request.

The author stays in control of timing. GitLab says you can finish iterating or address automated code review feedback first, then bring in a human reviewer when the change is ready. The feature is set up as a flow trigger in a project rather than as a project setting, so it sits alongside other agentic workflows a team already runs.

Why humans still review

GitLab positions the feature against a backdrop of growing code volume. The company cites a Reuters report from August 2026 saying that code changes on Meta's internal platforms grew 220% year over year, while changes that reached users as new features grew only 36%.

GitLab says its Code Review Flow already absorbs much of the review load. It reviews merge requests as they arrive, follows custom review instructions defined per project, and costs a flat 0.25 GitLab Credits per review ($0.25 at list price). Using an assumption of about 15 minutes of senior engineer time, or roughly $25 per manual review, GitLab claims automated review cuts per-review cost by 99%.

Automation does not remove people from the process, the company notes. Approval rules, CODEOWNERS and compliance policies still require a person to sign off on many changes, and for those merge requests, finding the right reviewer is often the first delay.

Works with other flows, and how to get it

GitLab says Recommend Reviewers Flow can be combined with other foundational flows in Duo Agent Platform:

• Code Review Flow, run first so automated feedback arrives before a person is brought in.
• Security Review Flow, which checks merge request changes for business logic vulnerabilities before they merge.
• Fix CI/CD Pipeline Flow, which diagnoses and repairs failed jobs so a broken pipeline does not hold up an approved change.
• Custom Flows, to build Recommend Reviewers Flow into an existing team process.

Intelligent reviewer suggestions are generally available in GitLab 19.5 for GitLab.com, GitLab Self-Managed and GitLab Dedicated. To use them, a team adds the Recommend Reviewers Flow trigger to a project and creates its next merge request. GitLab Premium and Ultimate subscribers can turn on Duo Agent Platform and use the GitLab Credits included with their subscription, and a free trial is available for new users.

What it means for teams

For teams that need human sign-off, the change targets the step before a review starts. According to GitLab, a better reviewer match means faster feedback, clearer accountability and a more balanced workload, with recommendations that come with visible reasoning that authors can override. Those benefits are the company's own claims, and the source gives no measured results for the feature itself.