Predictive Test Selection

An AI Co-pilot that correlates code-changes to tests to find failures up to 80% faster
80% of your tests are pointless—Launchable finds the tests that matter for a given set of code-changes

Quality a focus? Working with nightly, integration or UI tests?
Our AI can help.

Predictive Test Selection

Used by elite engineering teams

  • UKG Logo
  • BMW Logo
  • Sony Logo
  • Infosys Logo
  • Gocardless Logo
  • Optim Logo
  • Line Logo
  • Jenkins Logo
  • Vitess Logo
  • Delphix Logo
  • Keyloop Logo
  • UKG Logo
  • BMW Logo
  • Sony Logo
  • Infosys Logo
  • Gocardless Logo
  • Optim Logo
  • Line Logo
  • Jenkins Logo
  • Vitess Logo
  • Delphix Logo
  • Keyloop Logo
  • UKG Logo
  • BMW Logo
  • Sony Logo
  • Infosys Logo
  • Gocardless Logo
  • Optim Logo
  • Line Logo
  • Jenkins Logo
  • Vitess Logo
  • Delphix Logo
  • Keyloop Logo
  • UKG Logo
  • BMW Logo
  • Sony Logo
  • Infosys Logo
  • Gocardless Logo
  • Optim Logo
  • Line Logo
  • Jenkins Logo
  • Vitess Logo
  • Delphix Logo
  • Keyloop Logo

Your team drowns in a firehose of test failures before every release

We help them find calm amidst the chaos and ship with confidence.

Run

Rocket

Optimize tests to run

Filter

Filter

Filter away noise

Analyze

Dashboard and charts

Analyze the problem

Aware

Speaker

Raise awareness

Track

Chart and messages

Track the fix

SOC2 certified

It takes about one hour of a developer's time to set up Launchable

  • Four lines of changes in a CI Script
  • No developer support needed after setup
  • SOC 2 certified to keep your data safe
SOC 2 Type 1 Badge

Fast feedback for developers and optimized test execution costs

PTS curve - Run 50% tests to find 99% of failing builds

Find and run tests that are likely to fail based on code changes and other signals

A ML approach to radically improve test execution times & find failures.

  • Launchable learns your test suite to create a ML model (~2 weeks)
  • Your team chooses the best confidence for the test suite based on data from the ML model
  • Launchable predicts tests that will fail based on code changes based on the confidence target chosen.
  • Run tests up to 80% faster
  • Run an infrequent defensive full test run to catch any remaining issues.

Learn more about Predictive Test Selection

What is Predictive Test Selection video

Time Savings

Every time you run only a subset of your test suite using Predictive Test Selection, you save time.

MONTH2024-06Not final yet2024-05Final2024-04Final2024-03Final2024-02Final2024-01FinalPTS TEST SESSIONS134167148544099TOTAL DURATIONWITHOUT LAUNCHABLE1,787 hours2,228 hours1,598 hours203 hours203 hours1,609 hoursTOTAL DURATIONWITH LAUNCHABLE221 hours279 hours311 hours143 hours119 hours1,720 hoursTIME SAVED hours1,566(88%) hours1,949(87%) hours1,287(81%) hours60(30%) hours84(42%) hours111(07%)

Monthly Time Savings Report

A report that shows the amount of time saved by Launchable across test sessions.

Dry runs

Observation mode for dry runs

Run PTS in an “observation” mode to see how the model performs before turning it on in production.

Predictive Test Selection fits into your existing development pipeline

Unlock the ability to run a much smaller set of tests at various points in your software development lifecycle. With Launchable, tell your test runner exactly which tests to run based on the changes being tested:

The intuition behind Predictive Test Selection

Tests diagram

What if you could run failing tests first?

Test suites run randomly and are not optimized to find failures fast.

Reducing test feedback delay diagram

Reducing Test Feedback Delay

By running tests that are likely to fail first, we can reduce the time to first failure by a significant margin.

Risk-reward threshold diagram

Put a box around your risk-reward threshold

Launchable returns tests based on your thresholds.

Example: “Launchable give me all the tests that are likely to fail in 10 minutes” OR “Launchable give me all the tests that meet a 99% confidence threshold”

Works with your existing tools, languages, and processes

Results in weeks—no months-long DevOps transformations

Launchable's ML-based approach means it can work with existing languages and tools. Developers start seeing their dev cycles go faster without changing their processes.

Your co-pilot for Triaging, Understanding and Managing Test Failures

Quality a focus? Working with nightly, integration or UI tests?
Our AI can help.