Get 97% faster feedback with Smarter Testing by CircleCI

Fast feedback is the foundation of software delivery at scale. Long build and test cycles break developer focus, turning simple changes into momentum-killing pauses. Studies show that recovering from these interruptions can take twenty minutes or longer. Multiply that by dozens of commits per day and context‑switching time quickly turns into days of lost productivity.

This challenge is even more critical now that AI has increased both the volume of changes teams produce and the ease at which new tests are written. When new features can be created in seconds but test suites run for 5, 10, even 30 minutes, teams fall farther behind with every commit.

To keep up, you need validation that responds as quickly as your code changes. CircleCI is focused on removing the bottlenecks that slow down delivery and giving developers the fast, reliable feedback they need to stay productive. Smarter Testing is a major step toward that goal, reducing time to feedback by up to 97% while preserving the same quality standards you’ve always had.

Introducing Smarter Testing: Run only the tests that matter

Smarter Testing combines intelligent test selection with dynamic test splitting to tailor each test run to the exact changes in a commit. By examining coverage data and how code is modified, it determines which tests to execute and how to distribute them for maximum efficiency.

Here’s how it works:

• Coverage‑based test selection. When a developer pushes a commit, CircleCI analyzes code diffs, dependency graphs, and historical test behavior to identify which tests exercise the changed code. Only impacted tests and any new tests are selected for execution, preserving confidence while eliminating unnecessary runs.

• Dynamic splitting across parallel workers. Selected tests are then distributed optimally across available execution nodes in real time. This reduces the tail latency of long‑running tests and makes better use of compute resources.

Smarter Testing is language‑agnostic. It works with large test suites in common languages (JavaScript, TypeScript, Python, Java, Ruby) and is flexible enough for monolithic repositories and early monorepo implementations. Because it’s built into CircleCI, teams don’t need to manage separate services or maintain custom scripts. While other platforms may offer basic test splitting, Smarter Testing is the only unified, native solution for intelligent test selection and dynamic splitting at this scale.

Faster feedback, happier teams

We’ve been trialing Smarter Testing in private beta with several large customers who deliver a high volume of changes every day. These are complex, active codebases with substantial test suites, and the teams using Smarter Testing are seeing major reductions in the number of tests they run on each commit. Many projects skip more than 90 percent of their tests because those tests aren’t affected by the change.

These are meaningful gains. Early users report cutting their time to feedback by 90%, even as they continue to grow their codebases. In our own internal testing, Smarter Testing reduced suites from roughly ten minutes to around ten seconds, a reduction of more than 97%.

Improvements at this scale reshape how teams work. Developers can commit more often, keep a steady rhythm, and avoid the long pauses that break momentum. Pipelines benefit too. With far fewer unnecessary tests running, there are fewer flaky failures and false alarms, which makes the entire validation loop smoother and more predictable.

Accelerate what’s possible

Software teams are shipping more changes every day, and AI is accelerating that pace even further. More changes mean more code to validate, more tests to run, and more pressure on pipelines that weren’t designed for this level of throughput. Faster feedback keeps teams in flow, helps issues surface sooner, and supports the rapid iteration AI is designed to enable.

Smarter Testing is the latest step in our move toward autonomous validation. It builds on the work we’ve already started with Chunk, the autonomous agent that fixes flaky tests, repairs broken builds, and handles the CI/CD maintenance teams don’t have time for. Together, these features help shift CI/CD from a reactive process to a proactive one, keeping pipelines healthy and moving at AI speed without constant human effort.

Smarter Testing is available now in closed preview. To join the customers shaping the future of fast, focused testing, apply for early access through the waitlist.