If you’re one of the thousands of Awesome Admins who engage with Flow every day, you know that building automation is often just the first step. The real work, and sometimes the real frustration, begins when you need to verify that your logic holds up against real-world data and complex business rules.
We’re excited to introduce Test Mode, a new view for testing that lays the groundwork for our future testing capabilities and gives immediate benefits to the testing and debugging experience.
Flow has grown up. With over 690 billion monthly executions and growing, it’s no longer just a tool for simple tasks — it’s the engine powering mission-critical enterprise processes. However, the capabilities for testing your flows still have some gaps that prevent companies from adopting Flow even further.
Debugging and testing currently require separate actions. While you use “Debug” to troubleshoot errors or test flow changes, preventing those errors from recurring necessitates building an automated test. Automated tests transform manual test scenarios into permanent regression tests by using isolated test data and assertions (expected results) to ensure flow reliability. These automated tests must then be run either as part of deployments or run from the Flow UI. For these upcoming features to shine, a dedicated testing UI was essential.
We’re merging debugging and automated testing into a single, cohesive experience to simplify troubleshooting and test creation. Our goal is to remove the barriers to test automation to enable everybody to build reliable, robust flows through easier test creation, thereby reducing maintenance effort and freeing up IT teams to focus on innovation.
The first change you’ll notice is the removal of the Debug button from the toolbar. In its place, we’ve introduced a new Test Mode button, a dedicated space designed to make validating your work more intuitive.
To start using Test Mode, check the ‘Enable Flow Test Mode (Beta)’ box in the Process Automation Settings page.
Test Mode will be available in our beta program for:
When Test Mode launches, our plan is to support Screen Flows and Scheduled flows. Also note that by opting in, any target org that you deploy a test to will also need the beta feature enabled.
One of the most immediate benefits of Test Mode is the ability to save your flow runs as manual test scenarios.
With Test Mode, once you set up and run your flow, you can now optionally save and name the configuration for future use. This saves your exact configuration, including the triggering record, any input variables, and your mocked outputs.
This effectively creates a reusable test you can use later or share with coworkers. You can pull up a library of scenarios (for example, “New High Priority Case” versus “Existing Closed Case”) and run them with a single click.
Think of these manual tests as a bridge to test automation. A saved debug run is essentially an automated test without expected results. Once you save a scenario, you can easily add expected results, known as assertions, converting your manual troubleshooting session into a permanent regression test that protects your flow into the future.
Our goal is to ensure that all data types can be configured when testing or troubleshooting. With Test Mode, we’re introducing two new capabilities to elevate the test and troubleshoot experience with more on the way.
With Test Mode you can now mock, or simulate, the output of subflows and actions. With mocking, you can provide a user-defined value for the output of specific elements without actually executing the element. Need to test a flow that makes an HTTP callout without actually hitting the external API? You’ll be able to define a fake response so you can test the rest of your flow in isolation — a must-have for test automation and making troubleshooting easier for isolating issues.
Beginning in Winter ’27, with Test Mode you’ll be able to see how much of your flow is covered by automated tests with expected results. The canvas highlights the covered code paths making it easy to identify paths not covered by tests.
We’re transforming the testing and troubleshooting experience thanks to Agentforce. When logical collisions or runtime errors occur, you’re no longer limited to analyzing a flow in complete isolation. Instead, Agentforce pulls in the entire transaction’s context. It will synthesize Apex triggers, validation rules, user permissions, cumulative governor limits, and more, to pinpoint the exact root cause of failures across complex, chained automations. Backed by human-readable error translations and on-canvas help, troubleshooting cross automation issues that took hours will soon take minutes. Stay tuned for more!
This same intelligence will extend into your pre-deployment testing too, meaning you no longer have to manually seed volatile or tedious test data. Agentforce analyzes your flow’s metadata to automatically generate comprehensive test scenarios, mock element dependencies, and spin up isolated data silos complete with simulated edge cases and required input variables. What took days and hours to stress test your flow will also reduce to minutes.
In Winter ’27, Agentforce for Flow Headless will evaluate your flow’s logic and structure to draft a set of test scenarios in your console. Upon your approval, it automatically generates the tests, attaches them to the flow, validates them, and runs the execution.
To further improve your test reliability, Agentforce for Flow Headless can generate dedicated, isolated test data for your scenarios — completely separate from your actual org data. You can inspect this data anytime inside the Test Mode Scenario panel for each individual test.
Test Mode is a major step toward making Flow a better enterprise-grade platform. We can’t wait for you to try it out! Start by activating ‘Enable Flow Test Mode (Beta)’ in your Process Automation Settings. Then, open any record-triggered or autolaunched flow and select the Test tab to access the dedicated testing mode. From there, you can configure and save manual test scenarios, and eventually add assertions to convert your troubleshooting into permanent, automated regression tests. We look forward to your feedback as you build more reliable flows!
Get started with Agentforce Vibes Agentforce Vibes might seem like a strange starting point for the average Salesforce Admin, but it can be a very powerful tool. And it’s becoming more and more powerful as time goes on. To understand its potential for admins, it helps to start with the basics. When you open Vibes […]
Another year, another incredible Dreamforce- powered by you, amplified by agents! From connecting with other builders to getting sneak peeks of upcoming product releases, we are leaving energized and ready to build. Some of this year’s biggest moments came from admins bringing the Agentic Enterprise to life in the Admin Keynote. The keynote featured Radiant […]
It’s tempting after seeing an Agentforce demo to want to immediately start building an agent in your org. The reality is there’s a step before configuration: making the metadata make sense. Why? Metadata gives Agentforce the context and structure it needs to understand and interact with your data accurately. But reviewing all your metadata can […]
