AI Builder in workflows now more powerful
What's New
The AI Builder in workflows has been completely rebuilt from the ground up. The new version replaces the previous experience with a transparent, streaming, context aware copilot that shows you exactly what it's doing at every step. A new extensible architecture under the hood means every future capability ships faster and compounds on what's already live. Expect some really cool updates in coming weeks.
Live Streaming UI
Every build, edit, and query now streams results in real time with a step-by-step progress card. You can see exactly how your request is being processed — whether the AI is asking clarifying questions, searching for actions and triggers, creating the schema, placing it into the workflow, or validating the result.
Compound Requests in a Single Turn
Ask for multiple changes at once (e.g., "rename, add a 24h wait, turn off re-entry") and the AI executes them all in one go. No need to submit one request at a time.
Conversational Memory
The AI retains full session context. Say "actually make that 48 hours" and it knows exactly what you're referring to, no need to repeat yourself after every message.
Start fresh from the same canvas
If you trigger a build on a canvas that already has a workflow, the AI asks whether to start fresh or edit what's there. Earlier you would need to create a new workflow manually to start a new AI workflow creation.
Context Aware Empty States
The AI sidebar adapts to where you are. A blank canvas shows build-from-scratch prompts. An existing workflow shows "Make it / Fix it" suggestions based on your current flow. Chat mode surfaces knowledge prompts — one click to go.
Chat to Build in One Click
Ask a build question in Chat Mode and a "Continue in Build Mode" button appears, switching modes and auto-submitting your request instantly.
Session Continuity Across Panel Close
Close the sidebar, keep working, reopen it, and your full chat history and progress are right where you left them.
Learn Mode 2.0
Answers stream in real time with related follow-up questions surfaced at the end of every response, powered by a dedicated knowledge base search tool.
White Label Safe
The AI never references the platform by name in any output. This is enforced at the architecture level, not just filtered in responses. Agencies can expose the builder to their end customers with zero cleanup.
Why It Matters
This is the most significant upgrade to the Workflow AI Builder since launch. Every interaction is faster, more transparent, and more reliable. The extensible tool architecture means new capabilities no longer require a pipeline rewrite they plug in. What ships next builds on everything already live.
Conversation AI – Separate Controls for Bot Sleep on Manual vs Workflow Messages
What’s New
We are introducing more granular controls for how Conversation AI pauses (sleeps) when outbound messages are sent.
With this update, the setting “Send bot to sleep when I send a” has been split into two independent options:
- Manual Message
- Workflow Message
This gives users greater flexibility in controlling bot behavior based on how messages are triggered, improving automation precision and reducing unintended bot pauses.
By default, existing setups will continue to behave as they currently do, ensuring no disruption unless users modify these settings.
How it Works?
1. Default Behaviour (No Change)
- If no changes are made:
- The bot will continue to follow the previously configured sleep behavior.
- Existing automation and manual interactions remain unaffected.
2. Independent Sleep Controls (New)
- Users can now configure bot sleep behavior separately for:
- Manual Messages (messages sent directly by a user)
- Workflow Messages (messages sent via automations/workflows)
This allows:
- Sleeping the bot only for manual interactions
- Sleeping the bot only for automated messages
Or both (same as previous behavior
3. More Precise Automation Handling
With separated controls:
- Manual interventions can pause the bot without affecting workflows
- Automated messages can be sent without unintentionally stopping the bot
- Greater flexibility in hybrid (manual + automated) communication strategies
4. Reactivation Remains Configurable
- The “Reactivate bot after” setting continues to work as before
- Applies regardless of whether sleep was triggered by manual or workflow messages
Why this Matters
- Prevents unnecessary bot pauses during automated workflows
- Enables better control in mixed manual + automation use cases
- Improves conversation continuity and user experience
Gives users more granular and predictable bot behavior
Note
This is an enhancement to existing functionality.
Current behavior will remain unchanged unless users update the new settings.
Social Planner: Add Thumbnails in CSV Posting
We didn’t just improve bulk uploads, we gave you control over how your videos look before anyone clicks.With this update, CSV uploads now support custom video thumbnails, making bulk publishing more powerful and visually optimized.
🚀 What’s new
- Thumbnail Support in CSV Uploads: You can now add a thumbnailUrl field in your CSV/XLSX file to define a custom thumbnail for your video posts.
- Auto-Applied to Video Posts: The thumbnail is automatically applied to the first video in your post during upload, no manual intervention needed.
- Updated CSV/XLSX Templates: Sample templates now include the thumbnailUrl field, making it easy to get started with the correct format.
👷 How to use this feature
- Visit Marketing >Social Planner> Create Post> CSV Post
- Download the latest CSV/XLSX template
- Add your image link in the thumbnail Url column
- Upload your file as usual in Social Planner
- Your video post will automatically use the provided thumbnail
💡 Why it matters
- Improve click-through rates with better thumbnails
- Maintain brand consistency across bulk uploads
- Save time by managing thumbnails during scheduling
- Enhance visual quality of video content at scale
📝 Notes
- Thumbnail applies only to the first video in a post
- Multiple video thumbnails are not supported
- Behavior may vary across different social platforms
Snapshots for contact page customization
Following the initial labs release of contact page customization—which introduced up to five custom views per sub-account—we are now expanding its capabilities with snapshot support. Admins can seamlessly copy and reuse custom views across multiple sub-accounts without rebuilding them from scratch.
What’s new?
- Snapshot support for contact views: custom contact page views can now be included in snapshots and shared across sub-accounts.
– Cross sub-account copying: easily transfer custom views from one sub-account to another.
- Reduced manual setup: eliminate the need to recreate views individually in every sub-account.
- Scalable customization: maintain consistent layouts across multiple accounts with minimal effort.
How it works?
- Go to agency view and click on account snapshots.
- Create a new snapshot.
- Enter the snapshot name and select the sub-account from which you want to add views to the snapshot
- In contact detail views, you will find the custom views.
- Select the views you want to copy and click create.
- Push the update to the linked sub account
- The custom views should be ready
Why this matters?
Managing multiple sub-accounts previously required repetitive manual effort to recreate the same contact views. This update removes that friction by enabling reuse and standardization at scale. It saves time for admins while ensuring consistency in how teams view and manage contact data—making customization more powerful and efficient.
Smart Sending Ramp-Up for New Agencies on Shared Domains
Details:
We've introduced an intelligent, automated email sending ramp-up system for new agencies using AIOTB Shared or Agency Shared domains.
What's New
New sub-accounts no longer start with an uncapped daily sending limit from day one. Instead, they begin at a conservative threshold and automatically graduate to higher limits as they demonstrate healthy sending behavior — progressing through 8 defined stages without any manual action required.
Who Does This Apply To?
- New agencies (signing up on or after April 27, 2026) — Fully enrolled in the ramp-up system from day one. All shared-domain sub-accounts follow stage-based sending limits automatically.
- Existing agencies (created before April 27, 2026) — Completely unaffected. No changes to their current setup or limits.
- Migration plan for existing agencies — Will be shared by end of Q2 2026.
Key Highlights
- Automated Stage Progression (Stages 1–8) — The system evaluates bounce rates, spam complaints, AUP violations, and send volume on every check. Clean senders move up automatically; senders with issues move down. No manual action needed from agencies or support.
- Sending Limit UI on Email Services Page — Sub-account users on a shared domain now see a live Sending Limit panel showing Daily Limit / Sent Today / Remaining, their current stage and status badge, contextual alerts, and a full stage history timeline.
- Automatic Limit Controls — For agencies created from April 27, 2026 onwards, the manual "Update Limit" control is replaced with an informational banner. This applies to both the Agency UI and the Support Portal.
Why This Matters
Shared domain reputation is a collective asset. Previously, there was nothing stopping a new sender from arriving and damaging that reputation before anyone could react. This release shifts the platform from reactive to proactive — enforcing a trust-building period for every new sender, automatically and at scale, while giving users full transparency into their own limits and progression.
