Managing pipelines just got faster—and a lot more visual.

What’s New

• Create and edit pipelines directly from the pipeline dropdown
• Assign colors to each stage within a pipeline
• Choose how colors appear: Default (no color), Colored dot or Background tint
• Stage colors now reflect across key views like Kanban and record details

How It Works

• Open the pipeline dropdown in Opportunities
• Click + New pipeline to create instantly

• Use the edit icon next to any pipeline to modify it

• Create or Edit a pipeline to assign stage colors

• Select your preferred display style (dot or background tint)

Why This Matters

• Faster pipeline management without extra navigation
• Better visual clarity across stages and deals
• Easier to scan pipelines and prioritize work
• Consistent experience across views

Notes

• Color preferences apply at the pipeline level
• All users see the same stage color configuration

Book Services via Conversation AI

What’s New?

Users can now book appointments in Services Calendar directly through chat using Conversation AI across all supported channels, SMS, Live Chat, Facebook, Instagram, and WhatsApp.

Instead of filling out forms, customers can simply describe what they need in natural language, and the AI agent takes care of the booking.

This introduces a conversational-first approach to service bookings.

How It Works

Conversation AI is now fully integrated with Services Calendar configurations, supporting:

• Multiple services
• Location and staff selection
• Add-ons and variations

The AI understands user intent (even across multiple turns), maps it to the correct service setup, and creates accurate bookings.

You can also control:

• Which services the AI should manage
• Custom booking instructions for the agent

— This is a new feature altogether from scratch , based on the feedbacks lots of enhancements will be planned and released

—

How to Use

1. Create services in Calendar Settings
2. Enable the feature from Labs
3. Enable appointment booking in Prompt-based bots
4. Select the services you want the AI to handle

• private services are not supported. only the services which are public facing is supported

5. Start a conversation on any supported channel

—

Additional Capabilities

Full lifecycle support for service bookings:

• Book appointments
• Reschedule
• Cancel
• Send booking link instead (Optional)
• Pause bot after booking
• Trigger workflows
• Transfer conversations

—

Preview of Configuration

• Follows Existing Setup

Conversation AI strictly follows the configuration defined in your Services Calendar.

• Configuration Changes

Any updates (services, availability, staff, etc.) must be made at the Service Calendar level to reflect in AI bookings.

Payment Criteria

• Pay in Person

Appointment is booked and confirmed directly

• Online Payment

Appointment is created but remains in an unconfirmed state

Payment collection is not handled by the bot yet

Note: Conversation AI does not currently support collecting payments during the booking flow.

—

Key Innovations

• Memory Layer

Remembers user inputs even if they deviate mid-conversation

• Ambiguity Detection

Asks clarifying questions when needed

• Booking Preview

Shows details before confirmation

• Built for the Future

Scalable foundation for upcoming calendar types and advanced conversational booking

—

Coming Next

Support for Services V2 in Flow Builder bots is planned next.

—

Overview

We've shipped major reliability improvements to the Workflow AI Builder, drastically reducing generation errors. This release focuses on ensuring the AI Builder delivers a consistent, error-free experience at scale.

What's Fixed

1. Conditional Generation Errors – Eliminated

The AI Builder previously failed to generate branches or actions involving conditions (If/Else, Wait with Condition of Contact Reply, Decision Maker, Find Contact, etc.), causing entire workflow generations to fail.

• Before: Conditional generation errors occurred regularly during workflow creation
• After: 0% error rate
• Result: 100% elimination

2. Node Generation Failures – 10x Lower Failure Rate

Valid workflow nodes were failing to generate, causing entire workflows to fail during creation.

• Before: ~3.5% failure rate
• After: ~0.35% failure rate
• Result: 10x reduction in failure rate, despite 3x higher generation volume post-GA

Why This Matters

Our focus for the AI Builder is on three pillars: lower error rates, low latency, and correctness of generated workflow output. We prioritized fixing these errors before going GA, and the results reflect that investment. As adoption grows, we'll continue improving AI Builder reliability to ensure workflows generate correctly and consistently at scale.

Overview:

You can now measure how well Conversation SLAs are being met with the new SLA Performance Dashboard. While SLAs helped define response time expectations, there was no clear visibility into performance. This made it difficult to understand where delays were happening or how consistently SLAs were being met. This dashboard brings all of that into one place, giving you clear visibility into response times, SLA adherence, and overall performance trends.

👀 What’s New?

The SLA Performance Dashboard provides a comprehensive view of SLA adherence across messages, users, and channels. You can now see how many messages meet or breach SLAs, understand average response times, and track trends over time to identify patterns. The dashboard also breaks down performance at a user and channel level, making it easier to pinpoint where improvements are needed.

How to Use

• Go to Conversations → Analytics → SLA Performance and select a time range using the available filters.

• Select a time range and review the overview for overall SLA performance

• Use the trend chart dropdown to switch metrics

• Analyze user and channel reports. Use sorting and filters to drill into specific insights

Note

• Only messages with completed SLA evaluation (met or breached) are included in reports
• Conversations marked as read without a reply are excluded
• SLA performance is attributed to the assigned user at the time of evaluation
• Only Admins can access SLA Performance reports. Data for SLA Performance Dashboard is available starting March 23, 2026.

You can enable this feature from Labs – Conversations SLA Performance Dashboard. This feature is available only on the New Conversations interface. To enable the new interface, turn on CRM Redesign: Conversations and Contacts Page(Beta) from Labs.

We’ve introduced enhancements to improve email deliverability and platform security by automatically blocking emails sent to invalid or restricted domains.

What’s New

Automatic Blocking of Invalid Domains

Emails associated with known invalid or restricted domains are now blocked before sending.

Pre-Send Validation

Invalid email addresses are detected early and prevented from being used for email communication, reducing delivery failures and bounce rates.

Clear User Messaging

When an email is blocked due to an invalid domain, users will see the following message:

“This email address is invalid as the domain is restricted for security reasons.”

Why This Matters

• Enhances email deliverability and protects sender reputation
• Prevents failed deliveries caused by invalid or restricted domains
• Strengthens platform security and compliance
• Provides clear and actionable feedback to users