Summer ’25 is almost here to deliver powerful enhancements across Flow and Orchestration that help admins, developers, and business users build faster, smarter, and more scalable automations. Whether you’re deep into Flow Builder daily or just dipping your toes into automation, there’s something in this release for you. Let’s explore what’s new, along with use cases to help you see the value right away.
The process of starting a new flow just got a major UX upgrade. With the new “Create New Flow” modal, users get a streamlined, intelligent experience when launching into Flow Builder. You can access this directly from Setup as well as in the Automation App. This modal helps users identify the right flow type with smarter filtering, search capabilities, and different categories.
There’s also a section for your most frequently used flow types and templates, which is perfect if you find yourself creating the same kinds of flows regularly (looking at you, screen flows). This enhancement addresses a common problem for newer builders: picking the wrong flow type and realizing it halfway through building. By making it faster and more intuitive to choose the right type of flow from the start, this modal helps reduce errors, boost confidence, and save time without requiring a single change to user permissions.
Reusing logic is a hallmark of good development, so we’re making it easier than ever to templatize your Flow work. The new “Save as Template” option is built right into the standard save flow dropdown, making it simple to create reusable blueprints for flows you build often. Previously, creating a template meant digging into Advanced Settings and manually checking a box. We heard you—that it was easy to miss because it was not exactly discoverable. Now, you can click the arrow next to “Save as New Version”, choose Save as Template, and fill in a name, API name, and description.
This functionality supports all flow types that allow templating, giving teams a low-friction way to build a library of consistent, high-quality flows. Whether you’re an architect trying to drive best practices or an admin supporting multiple business units, templates make it easy to share logic without reinventing the wheel.
Flow Builder’s canvas just got a lot more responsive, thanks to single-click editing for element cards. Previously, editing a Flow element meant a double-click or, worse, clicking the wrong thing and getting distracted by another pop-up. With this enhancement, a single click opens an element’s configuration panel, speeding up the most common builder task: editing your logic.
Additional updates include a new “⋮” menu on each card, where you can cut, copy, delete, or assign a fault path, all without navigating to another part of the canvas. Element descriptions now appear on hover, so your screen isn’t cluttered with text boxes unless you need them. Plus, the new “sticky zoom” feature preserves your zoom level when selecting or editing an element, so you’re not constantly reorienting yourself. These changes don’t just make Flow faster—they make it feel modern and intuitive. Small touches, big productivity gains.
Summer ’25 introduces zooming behavior that aligns with what users expect from other modern interfaces. You can now zoom in and out of the Flow Builder canvas using standard keyboard shortcuts (Cmd/Ctrl + Option + Plus/Minus. And for trackpad users, pinch-to-zoom finally works.
This might sound small, but for anyone working on large or complex flows, constantly reorienting your view is part of the job. The new zoom model is smoother, less disruptive, and uses muscle memory you’ve already built in other apps. It’s particularly helpful for laptop users who rely on gestures to navigate. When paired with sticky zoom, it creates a more stable and satisfying builder experience overall.
Finding deeply nested fields in complex flows has traditionally been a pain—it’s like playing hide-and-seek with your data. But with the updated Resource Picker, you can now search up to 10 levels deep, using the new Expand Search mode (currently in beta).
This is especially useful when working with structured objects or external data that has multiple layers—think Data Graphs, nested collections, or MuleSoft-connected data. For example, if you’re trying to find the Close Date field on an Opportunity, the expanded search will help you pinpoint it in seconds. No more clicking blindly through dropdowns, or writing down field paths on a sticky note. This feature brings serious depth and efficiency to one of Flow Builder’s most-used tools.
Picklists and dropdowns just got more user-friendly. Previously, selected options in the new Resource Picker would often display their API names, which was especially confusing if you’re working with external systems where labels and values can be significantly different.
In Summer ’25, picklist values use the user-facing label, not the backend API name. This format improves visual clarity, reduces cognitive load, and makes it far easier to review your configurations at a glance. This is particularly helpful when configuring connectors like Jira or MuleSoft, where picklist values can sometimes look like alphabet soup. Logic still runs on API names behind the scenes, but you get a clean, understandable UI to work with. It’s a simple change that removes a lot of confusion and increases flow readability for anyone reviewing your setup.
Time is no longer an afterthought in Flow Builder. With this beta enhancement, the Time data type is now supported across Flow elements, including the Resource Picker, Formula Builder, and Expression Editor. You can create time-based automations, apply logic to specific hours or minutes, and even configure time inputs with a new dedicated Time Picker.
For example, let’s say you want to run automations tied to operating hours, appointment slots, or delivery windows. You can now fetch time values directly, compare them with other time-based criteria, and use them just like you would with Date or DateTime data types. This unlocks more precise scheduling logic and reduces the need for workarounds involving strings or complex formulas. It’s especially helpful in scenarios like call center workflows or SLAs where the hour matters just as much as the day.
Say goodbye to the spaghetti of nested Get Records elements and loops just to gather related data. With Summer ’25, Flow introduces a new option inside the Get Records element that allows you to pull related records from multiple objects—all in one step. This means you can retrieve an Account, its related Opportunities, and those Opportunities’ Line Items without breaking a sweat. No more writing multiple queries or worrying about hitting limits.
The configuration interface allows you to specify which relationships to follow and which fields to retrieve, giving you control without overwhelming complexity. If you work on data-rich UIs or logic-heavy automations (wink, wink, looking at you, screen flows for quote generation), this dramatically reduces development time and cognitive overhead. It’s also a win for performance and org hygiene, since fewer elements mean cleaner flows with less room for errors.
Flows can now invoke AI agents, finally enabling two-way collaboration between deterministic and generative logic. Previously, agents could run flows, but flows couldn’t call agents. That changes now with the new Run Agent action, available across all flow types once your agent is activated in Agent Builder.
Here’s an example use case: a Deal Desk flow that can automatically trigger an agent to summarize a request, categorize it, and draft a response for a human to review. This new functionality allows customers to blend traditional automation with AI-powered reasoning, all inside the same flow. The only caveats? You’ll need Einstein credits, and it’s limited to Agentforce Default and Service Agents for now. Still, this is a major step toward making flows smarter and more autonomous.
Sending emails from flows has always been powerful, but the configuration experience? Not so much. In Summer ’25, the Send Email action gets much-deserved love, giving you a much better experience. You can configure inputs more easily and select recipients, sender addresses, and attachments using a modernized input panel. You’ll also get a hierarchy-based layout that makes it easier to see what you’ve configured and what’s still needed.
Multi-select and single-select inputs are supported, giving users more control over complex email logic. One standout improvement is how easy it is to define the sender, which was previously a hidden setting that required too many clicks. Whether you’re building onboarding emails, case notifications, or internal alerts, this update makes the whole process smoother and less error-prone. Consider it a small but mighty upgrade to one of Flow’s most-used (and most misunderstood) actions.
Have you ever added a subflow or invocable action and wondered, “What outputs does this actually give me?” You’re not alone. In Summer ’25, we’re adding a new “View Output Resources” section to the configuration panel of Action and Subflow elements. This section shows each output’s name, label, data type, and a short description, where available. It’s a huge improvement over the old model, where you had to know the outputs in advance (or just guess and check).
Now, the information is right there when you need it, which means fewer trips to documentation and fewer mistakes in using those outputs downstream. Hovering over the information icon gives you even more context, making this feature especially valuable when using third-party invocable actions, prebuilt subflows, or newly released packaged components. It looks like a small feature, but it’s one that dramatically improves flow readability and builder confidence.
Let’s talk about one of the most common trip-ups for flow builders: callouts in record-triggered flows. Previously, if you put an external callout in the immediate path, Flow Builder would stop you—but not until you try to save it, and that’s way too late. Summer ’25 improves the experience by adding clearer messaging and a more visible toggle in the Start element for configuring asynchronous paths. If you add a callout action to the immediate path, you’ll see a warning right away.
This small change makes a big difference by increasing the visibility and discoverability of best practices around async logic. Importantly, the underlying functionality hasn’t changed, just the in-app guidance. The result is fewer save-time errors, better flow structure, and a much smoother development experience.
Previously, if you wanted dynamic behavior on a screen, like showing data based on a user’s input, you had to rely on Action Buttons or build custom Lightning web components. In the previous release, we introduced Reactive Screen Actions: a game-changing capability that lets screen flows react to user input changes instantly, without requiring a button click or manual refresh. In Summer ’25, we’re making it available for everyone.
Imagine a user selecting an account from a lookup field. A related Data Table on the same screen can now show an account’s open cases in real time or calculate totals instantly when a user selects line items in a quote flow. What once needed custom automation in Apex or AppExchange apps can now be built declaratively thanks to this powerful addition. It’s especially useful for record-based dependent picklists, data rollups, duplicate record validation, and more. With this GA release, Flow becomes a truly reactive UI platform—and the need for screen refresh workarounds is finally behind us.
Reactive logic is powerful, but sometimes too much reactivity becomes a problem—especially when an action fires every time any input changes, even when it doesn’t need to. To solve this, we’re introducing an option to set conditions for when a screen action should run. For instance, maybe you only want to fetch records when a particular picklist has a certain value. Or perhaps you want to suppress actions when the screen is revisited. With these new settings, you can fine-tune when and how screen actions execute, making flows faster, more predictable, and more efficient. This is especially helpful for large flows where performance is a concern or for mobile users where latency matters. It’s a small checkbox-level change that delivers big control to flow designers.
Anyone who’s ever tried to make multiple fields line up just right in Flow Builder knows the pain of working with the old Section-based layout system. While Sections supported multiple columns, they didn’t offer per-component width control, leading to annoying misalignments and strange whitespace.
Summer ’25 fixes that with the introduction of 12-column grid-based layout control at the individual component level. Now, you can set each component’s width explicitly, and they’ll sit neatly next to each other in perfect rows. You can even set widths dynamically in repeaters. For example, you might want three fields side by side in a wide desktop layout, or a single full-width component on mobile. Vertical alignment control is also included, so fields stack and align exactly how you want. This enhancement puts true design flexibility in the hands of flow creators—no CSS or code required.
What looks great on desktop may feel like a mess in a mobile app or utility bar. Previously, the only way to test your flow’s responsiveness was to activate it and then run it in the target environment. With the new screen preview feature, you can simulate different screen sizes right inside Flow Builder. Whether you’re building for small utility bar layouts or mobile devices, just switch the preview size and see how your screen responds in real time. This adds a crucial layer of confidence to screen flow development and reduces rework caused by layout surprises. This change makes designing and validating responsive flows significantly easier, especially for admins supporting service teams (who often work in narrow console environments) or mobile-first field users.
Sometimes, your users don’t want to read, they want to see. The new Visual Picker component brings icon-based decision-making into Flow, letting users select options represented by SLDS icons in a tile layout. For example, you could display a “Thumbs Up” and “Thumbs Down” to capture yes/no responses, or show icons for product categories like laptops, monitors, and accessories. The component is configured using Choice resources, and for each choice, you can select an icon and customize how it appears. While it doesn’t yet support Record Choice Sets or Picklist Choice Sets, it shines in guided selling, surveys, and user onboarding flows where visual affordance improves usability. It’s one of those features that makes a screen feel more like an app and less like a form—and your users will notice the difference.
Speaking of visual polish, Choice Lookup components now support icons too. This small, but mighty, enhancement allows you to assign icons to individual choices, making dropdowns and lookups more informative and user-friendly. For instance, imagine a list of issue categories in a support form and giving each of them their own visual cue. This not only helps users make selections more confidently but also improves accessibility and differentiation between similar-sounding options. Similar to the Visual Picker, icon support is limited to standard choice resources, but it’s a welcome upgrade for anyone designing high-clarity, low-friction user interfaces. It may not change what your flow does, but it definitely changes how it feels.
We all know that in a hectic world, business rules can change the moment after approval is submitted. When that happens, users often need to pull their submission back. Recalling an approval simply killed the process without any follow-up actions. This led to silent failures, confusion for downstream teams, and leftover data that had to be cleaned up manually.
To address this, we’re adding a new Recall Path feature that gives admins the ability to define post-recall automation. When someone recalls an approval request, the flow can automatically trigger cleanup actions, send notifications, and reset related statuses. For example, recalling an expense report might notify the accounting department and release reserved budget allocations. The beauty here is in the flexibility. You can define exactly what happens based on your business rules. It turns “Oops, never mind” into a structured, auditable event without adding complexity to your process.
Vacations, sick days, and back-to-back meetings happen, but approvals still need to move forward. With Flow Approval Process Delegation, Salesforce now supports native routing of approval requests to a delegate when the original approver is unavailable. This leverages the existing delegation settings in your org, so you don’t need to configure anything new. When enabled, work items are created for both the primary approver and their delegate, ensuring that approvals don’t stall while waiting for someone to return. This eliminates the awkward workarounds many orgs rely on, like mass reassignment or Slack spamming, and helps keep your business processes moving without skipping a beat. It’s a quality-of-life improvement that also happens to prevent expensive process bottlenecks.
One of the biggest gaps in flow-based automation has finally been filled: You can now submit a record for approval directly from a flow. This unlocks a huge number of use cases where preapproval logic needs to be evaluated before routing a request for signoff. Let’s say you’re building a customer onboarding flow. Based on risk score and compliance criteria, you may want to submit a KYC approval request—but only if certain thresholds are met. With this feature, you can trigger the approval process from within the flow itself, making it seamless for the end user. The action supports the default one-approver process out of the box, but it can also integrate with more complex hierarchies you’ve built. It effectively bridges the two automation worlds: Flow logic and formal Approval Processes, giving admins tighter control and a better user experience.
Let’s face it: Things break. APIs go down, payments fail, someone forgets to assign a required field. The proper risk management isn’t avoiding failure, it’s planning for it. With the new Fault Path capability in Flow Orchestration, admins can define what should happen when something goes wrong during an orchestration step. Previously, a failed step would stop the entire orchestration cold, often leaving records in a half-finished state and requiring manual intervention.
Now, you can create smart recovery paths that kick in automatically. For example, if a payment authorization step fails, you could retry using a backup method or send a Work Guide notification to the finance team. This enhancement brings orchestration closer to enterprise-grade process automation by letting you define not just the happy path, but the recovery plan too. It improves reliability, reduces operational cost, and gives users confidence that even if something fails, the system has a plan.
Flow Orchestration continues to grow as a global tool, so as a logical step, we added an option for admins to translate orchestration stage names using the Translation Workbench, allowing the Work Guide UI to display in the user’s preferred language. This is particularly important for multinational orgs where users work in different locales and expect to see UI labels in their native tongue. It’s a small detail with a big impact on user comfort and comprehension, especially for customer-facing or compliance-heavy workflows.
Say hello to a new flow sheriff in town: activation-triggered flows. Until now, Data Cloud Activations could send data to prebuilt connectors, but they couldn’t initiate complex logic or call custom systems. That changes in Summer ’25. Now, when a Data Cloud activation is triggered, it can launch a flow to take action. These are full-fledged flows, meaning you can run invocable actions, perform condition checks, or even call out to external APIs using HTTP callouts. For instance, a fashion company might trigger an SMS campaign when a loyalty customer hits a spending milestone. This feature brings the full power of Flow to Data Cloud activations, helping teams create highly personalized, cross-channel experiences. It’s low-code orchestration that thinks globally and acts instantly.
What happens when a flow fails at 3:00 AM? If you’re lucky, someone finds out by noon. If you’re not, it fails silently for days. That’s where Flow Logging via Data Cloud comes in. It’s a major step toward enterprise-grade monitoring and troubleshooting for Flow. Once enabled, this feature logs each flow run to Data Cloud, where admins and analysts can view execution details, track errors, and measure performance over time.
Setup is straightforward: Click the Install and Setup Logging button under Flow advanced properties, toggle logging in Flow Builder, and then analyze logs directly in Data Cloud. This kind of visibility is a game-changer for large-scale orgs that need to ensure reliability at volume. Heads up: Logging consumes Data Cloud credits and is only available for autolaunched and scheduled flows.
Flow debugging just got a serious upgrade. In the past, the debug panel was functional but cluttered, making it hard to find the data points that mattered during testing. Now, each debug item is displayed in a card format with a summary line, clearly showing what happened at each step. You can also filter and search through results, which is a lifesaver in larger flows. Want to find just the variables that changed? Or see which assignment step failed? No more scrolling endlessly through a wall of output.
The debug panel also resizes to 80% of the screen, giving you more room to view context without squinting or flipping tabs. If that weren’t enough, you can now copy the full debug log to your clipboard, allowing for external analysis, comparison, or even pasting into documentation. This enhancement turns debugging from a chore into a genuinely productive and efficient step in your build process. This new experience applies to flows that use the on-canvas debug experience (not available for screen flows).
Automated testing in Flow is no longer just a button-click operation, it’s now command-line friendly. With the new sf flow run test command, you can run Flow Tests directly from your terminal or CI/CD pipeline. This means you can test every flow in your org (or just specific ones) and get results in a structured format. You can see the test status in your CLI output, review them in Flow Builder, or query FlowTestResult records directly for more detail.
The best part? It also supports error path testing, not just happy paths. This brings Flow Builder closer to the DevOps ecosystem, enabling test automation during deployments and helping teams build robust, reliable, and regression-proof flows. Whether you’re doing full release testing or just running validation scripts before committing code, this enhancement closes a critical DevOps gap.
Here’s a gift for every developer who’s ever tried to do a diff between two flow versions in Git and thought, “Why is this showing up as changed when nothing changed?” The answer was always those X/Y layout references, which got regenerated even if the flow logic stayed the same. Time to celebrate—that noise is now gone!
When you save a flow in Auto-Layout mode, Salesforce will automatically strip out obsolete layout references from the XML. This means version comparisons in Git (or any source control tool) will now highlight only the real changes. It makes reviewing changes, tracking regressions, and doing code reviews much simpler and far more accurate. This may seem like a small behind-the-scenes update, but for teams practicing source-driven development, it’s a huge quality of life improvement.
The Summer ’25 release is a milestone for Salesforce Flow and Flow Orchestration. From AI-generated flows and deeply reactive screens to fault-tolerant orchestrations and CLI-driven testing, this release meets teams where they are—whether you’re automating your first screen flow or managing an enterprise-scale automation program. It’s all about speed, clarity, and confidence: building smarter, debugging faster, and delivering better business outcomes without unnecessary complexity.
What’s next?
We can’t wait to see what you build with all this new power at your fingertips. As always, join the conversation in the Salesforce Automation Trailblazer Community or reach out to your customer success team to go deeper.
Each release brings tons of amazing, new functionality and it can be a lot to digest. To help you make the most of Spring ’25, be sure to explore our Be Release Ready page, where you’ll find valuable resources for Salesforce Admins. Bookmark the page and check back for any updates as we continue to provide helpful insights!
Spring ’25 is almost here! Learn more about Flow features and check out Be Release Ready to discover more resources to help you prepare for Spring ’25. As usual, there are a TON of new features and improvements coming to Flow in Spring ’25. To help get you up to speed and excited, this post […]
#AwesomeAdmin Melissa Hill Dees wondered aloud to the Trailblazer community on Twitter how she could duplicate a Salesforce record while updating some of the fields using only declarative tools. Melissa says, when a record meets a specific criteria, “I want to duplicate the record, adjust some values, and save the new record.” This made me […]
Summer, by far, is one of my favorite seasons. It’s the time of year where the chances of me being cold are really slim. When I think of summers in the US, I think of lemonade, roasted corn on the cob, barbecuing all the things, sitting on the beach in P-Town with Mochi (my 6-year-old […]
