This update introduces a significant upgrade to the agent-tasking workflow, transforming task management from implicit bookkeeping to explicit workflow primitive. The core advantage is a reliable, flexible, UI-friendly system for managing multi-step agent tasks, along with improved planning and context awareness.

Tasks were managed using numeric indices, which were fragile and prone to errors.

Task transitions inferred, leading to potential mismatches and inefficiencies.

UI representation was limited with static labels and ambiguous statuses.

Introduced todo to handle task creation, transition and addition explicitly.

Replaced numeric task indexing with label-based transitions for clarity and accuracy.

Enhanced UI with dynamic status for active and completed tasks.

Improved planning and context awareness, allowing the model to better understand task sequences and dependencies.

Added descriptive labels for tasks, improving frontend progress display.

Maintained backward compatibility for existing UI while enabling richer labels- Technical

Centralized task management in a meta-tool, reducing prompt/tool overhead. Tool schemas for more structured outputs.

Updated system prompts to align with the new workflow, ensuring better agent behavior.

Enhanced context awareness, providing AI models with richer task-related information for higher-quality results.

More Reliable Progress Tracking: Explicit task transitions enhance dependability in multi-step jobs. Allows for dynamic task addition and accurate status updates. -Enhanced UI/UX:** Provides clearer progress states with human-readable labels.

Reduced Overhead: Streamlined architecture and schema handling lower resource usage and error rates.

Better Planning and Context Awareness: Enhances the model's ability to understand task sequences and dependencies, leading to more efficient execution.

We've added two in context safeguards to the Contact Tag trigger to prevent the most common configuration mistakes and protect users from unexpected workflow behavior.

When a tag is added to the Contact Tag trigger, a subtle info banner now surfaces to clarify that the trigger only applies to tags added after the workflow is published. The banner links directly to the retroactive enrollment guide, where users can follow the steps to bulk import existing tagged contacts into the workflow.

If a user attempts to save the Contact Tag trigger without configuring any filters, a warning now surfaces explaining that the workflow will trigger for every tag added or removed across all Contacts, which may cause unexpected behavior. The warning offers two clear paths forward:

You can now view, edit, and manage your website’s code right inside the builder. No switching tools. No extra steps. Just fast, real-time updates.

CodeMirror-powered editor with support for TypeScript, JSX, CSS, JSON, and more. Tabs, keyboard shortcuts, and unsaved-change indicators included.

Cmd/Ctrl + F for fast, VS Code-style search with case-sensitive and whole-word options.

Search your entire project, jump to results instantly with highlighted matches and line numbers.

Save and see changes live in seconds, no deploy step needed.

Errors show up immediately. Click “Try to fix” and let AI handle it.

New routes in code? They appear automatically in the route selector.

Every edit is saved. Preview, bookmark, or roll back anytime.

Get warned before leaving unsaved code changed, no lost work.

The Drip action has been upgraded with a set of transparency and reliability improvements, giving users full visibility into how their drip schedules will run and ensuring pacing is preserved across workflow state changes.

A live schedule preview is now available inside the Drip action during configuration. Users can see exactly when each batch is projected to run before publishing.

The statistics icon on the Drip action has been redesigned as a proper, discoverable button. Clicking it opens a detailed view of everything happening inside the drip.

When batch size is updated on a previously published workflow with contacts already queued, a soft blue informational note is shown:

"Note: Contacts already queued in this drip will use the previous batch size. New contacts will use the updated batch size."

This appears only when the workflow has been published at least once and contacts are actually present in the Drip action, keeping the message relevant and non-intrusive.

When a workflow moves from Published to Draft with contacts queued in a Drip action, those contacts are now paused for the full duration the workflow remains in Draft. Once republished, the drip resumes from where it left off instead of bursting all queued contacts out at once, preserving the original pacing and protecting sender reputation.

A tooltip now appears on hover of the Drip action (similar to the Wait step), allowing users to read the full configuration without opening the action. Examples:

Tooltips on the individual fields inside the Drip action have been refreshed with clearer descriptions and examples, so users can understand what each field does directly in context.

Drip schedules previously behaved in ways users couldn't easily explain, driven by three recurring gaps: 1)Active Workflow Time Windows silently shifting batch times

2)Mid-run setting changes applying only to new contacts without clear indication.

3)Workflows toggling between Published and Draft causing queued contacts to burst out all at once.

This release closes those gaps by surfacing constraints directly in the product, preserving drip pacing across publish state changes, and delivering reduction in guesswork around drip scheduling.

We’re excited to introduce Nested Folders, a powerful new way to organize your assets across Campaigns, Templates, and Sequences.

✅

You can now create subfolders within folders, allowing for a more structured and scalable organization system. Whether you're managing multiple clients, campaigns, or content types, nested folders make it easier to keep everything neatly arranged.

🔐

🎯

💡This release supports more than 5 layers of nesting as well for better organisation.

Earlier this month, we rolled out contact page customization to a select group of 100 customers in private beta (March 30th). The response? Extremely positive — with zero bugs reported and valuable feedback from users who had been asking for more flexibility in how their contact data is structured and viewed.

Based on the feedback and the customer calls we did for the new contact UI, it became clear that enabling customization directly solves a key customer problem. Building on this momentum, we’re excited to announce that

Previously, all users had to work with a single, fixed layout — regardless of their role or workflow. With this update, you can now:

This is just the beginning of making your CRM truly adaptable to your business.

Overview:

Hosting events just got a major upgrade.

You can now host events directly inside Communities with a new Live Room option.

Creating an event now automatically sets up your meeting:

Attendees now get a smoother experience:

We’ve improved the Add Task workflow action so you can set due dates with much more flexibility. You can now configure task due dates using days, weeks, months, or years, add a specific due time, see a live preview of the evaluated due date, and create task descriptions with rich text formatting.

In the Add Task action, the due date is now configured through a Due In input with a numeric value and time unit. You can optionally set a specific time, and the action shows a live preview of what the final due date and time would be if the workflow ran right now. Task descriptions now support rich text and keep their formatting when saved and displayed in Tasks.

Earlier the task action was limited with due dates to a very short window, which made it hard to plan follow-up work beyond 5 days or assign tasks for a specific time of day. This update gives you more control over long-term task planning and helps you create clearer, more detailed tasks directly from workflows.

You can now receive notifications for new messages in conversations you follow, ensuring you never miss important updates.

👀

⭐

Previously, only conversation owners could receive notifications for new messages. Followers had to manually check conversations, which could lead to missed updates. With this update, followers can stay informed in real time, improving response efficiency.

We've launched the AI Agent action in Workflow AI, bringing autonomous reasoning and decision-making into workflows. Until now, workflows were purely deterministic , every decision had to be pre defined as a branch, every condition anticipated, every field manually mapped.

The AI Agent changes that. Describe your goal, give the agent access to tools, and it plans the sequence, decides the inputs, handles edge cases, and executes all within a single workflow step.

The agent automatically searches and pulls context from across the entire CRM — contact details, pipeline data, calendar availability, opportunity history, and custom fields. No manual variable mapping or lookup steps required.

Skip rigid dropdowns and field-by-field configuration. Write instructions in your own words — say "assign the deal to Sarah" and the agent resolves "Sarah" to the actual CRM user. Reference "the Enterprise pipeline" and it identifies the correct one. Instructions can range from a quick paragraph to a detailed multi-page playbook.

Write rough, casual instructions and click Enhance Prompt. The AI restructures your input into a well-organized prompt with numbered steps, edge case handling, and messaging guidelines — turning a one line note into a production ready prompt.

Pre built agent configurations for the highest-impact use cases, each prefilled with instructions and tools. One click to add the required trigger, one click to apply, and everything remains fully editable:

Each tool includes a Let AI decide all field values toggle. Turn it on for full autonomy at runtime, or turn it off to lock specific fields (e.g., always assign to a particular user) while letting the agent decide the rest. Individual fields can also be toggled independently, giving you a full spectrum from fully autonomous to semi-autonomous configurations.

When enabled, the agent retains a rolling summary of past executions for the same contact. A lead pipeline tracker, for example, remembers what it sent last week and adjusts its messaging accordingly — evolving from a one-shot decision maker into an ongoing relationship manager.

The agent can return results as Text or JSON with a user-defined schema (String, Number, Boolean properties). Output feeds directly into If/Else branches, field updates, or downstream workflow steps. For example, the agent can qualify a lead and return {"qualified": true, "score": 85} for the workflow to branch on.

Every run produces a detailed trace in Execution Logs showing every LLM reasoning step, every tool call (with exact inputs and outputs), timestamps, token counts, and success/failure status. Toggle between table view and raw JSON to make prompt refinement a tight feedback loop.

Choose between GPT-5.2 (Low/Medium/High thinking), GPT-5.1 (Low/Medium/High thinking), and GPT-5 Nano (Fastest). GPT-5.2 Low thinking is the recommended default for most use cases.

AI Agent pricing is transparent and usage-based. Each execution costs:

Standard CRM tools (Send SMS, Update Contact, Add Tag, etc.) don't add any premium charge.

The AI Agent transforms Workflows from a deterministic automation platform into one where a true CRM agent can evaluate context, make judgment calls, and execute actions on its own — all within the workflow system you already know. Instead of mapping every branch and condition by hand, you describe the outcome you want and let the agent handle the reasoning, freeing you to build smarter automations in a fraction of the time.