June Releases

One-Time Password (OTP) Settings and Audit Log

A new One-Time Password page is available under Settings → Security and Compliance. Workflows can send a one-time code to a Customer's email or phone number to verify they control that address before taking action – such as looking up or canceling an order. A starter Workflow template (User address verification) is available to get up and running quickly.

The settings page lets Admins configure the sender brand name and email ‘From’ address. A built-in Audit Log tracks every send, delivery, and verification attempt by recipient.

Searchable Workflow Triggers with Version and Direct Links

Dropdowns in the "Before First Message" and "Before Handoff" sections of the Workflow Triggers tile now include a search box. Each listed Workflow displays its version, is searchable by name or description, and can be opened in a new tab via the link icon.

Reorderable ‘Else If’ Branches in Workflow Editor

Rearrange ‘Else If’ branches by using the "Move Up" and "Move Down" options within the editor instead of deleting and re-adding them in the desired order.

Workflow Step Cards in Conversation Review Show More Details

Workflow cards in the Inspect Summary surface more context for each step, making it easier for Admins to debug Workflow behavior and understand how AI decisions were made:

• Failed Workflow steps consistently show a red error banner on the title row across all step types. Expanding the card reveals the underlying error message.

• Recall steps show more detail about what was returned and whether the runtime fell back to a default value.

• Ask steps display the initial Workflow prompt, the Agent's outgoing message, the Customer's reply, and the collected value on one expandable row – each labeled and truncated to a single line for quick scanning, with full detail one click away.

• Use steps show the action name in a single collapsed row. Expanding reveals the raw input and output exactly as recorded, making it easier to debug what an action was called with and what it returned.

• Stop steps now lead with the final action the Workflow took. For "Go to" handoffs, the target Gladly Agent name is a clickable link in the row title. The expanded view shows reasoning, any Guides the Workflow redirected to, or errors.

Updates to Shopify Compliance Redaction Requests

Compliance Admins have more visibility, more export options, and more control over how they manage redaction requests:

• The daily digest email now shows when its counts were generated, the age of the oldest pending request, and the age range of pending requests per type – so Admins can distinguish a stalled backlog from a fresh queue without leaving the email.

• Compliance Admins can export the full redaction requests list as a CSV from the Redaction Requests page. The export includes the full Shopify Customer ID and surrounding metadata (request type, status, shop domain, fulfilment info), and respects any filters applied on the page.

• Compliance Admins can opt out of the daily pending redaction requests digest email by clicking the ‘Settings’ button within Security and Compliance → Redaction Requests and unchecking the email opt-in.

Audit Log Includes Agent Email and Name

Audit Log entries initiated by an Agent now display the Agent's email and name, making it easier to identify who made each configuration change without looking up the Agent ID separately.

View AI Conversation from the Conversation Timeline Summary

Admins and Managers can jump to the corresponding Gladly AI Conversation directly from the AI summary within the Conversation Timeline. Click the external link button on the summary card to open the Conversation Review panel for that AI response – no need to go through the overflow menu.

Granular Topic Requirements on Conversation Closure

Admins can now configure specific Topic requirements that must be applied before a Conversation can be closed. Some of these requirements might include:

• Require a topic from one or more parent groups.

• Require a minimum number of topics overall.

• Require specific freeform topics to be filled in.

Topics can also be marked as "Category only" so they act as organizing groups in the hierarchy and are never applied to a Conversation directly. When closure is blocked, the Apply Topics tooltip tells the Agent exactly which group or Freeform Topic still needs to be addressed.

Mid-Conversation Guide Reclassifier Now Default Behavior

Gladly AI now reassesses which Guide to use after every Customer message – not just the first one. If the AI initially picks the wrong Guide, or a Customer shifts topics mid-Conversation, Gladly AI seamlessly moves to the correct Guide. This behavior is now active by default across all organizations with no configuration needed.

A few tips to get the most out of this feature:

• Write clear Guide descriptions so the reclassifier can route accurately.

• Minimize overlap between Guides.

• Keep each Guide self-contained rather than redirecting to another.

Learn more about best practices for accurate Guide selection.

Export Rules Available via the Support Tool Portal

The Configuration Exports panel within the Self-Serve API Export Tool includes an Export Rules option that downloads your routing and automation Rules as a CSV.

Each rule includes its name, priority, trigger, conditions, and actions – with all IDs resolved to readable names (Topics, Agents, Inboxes, etc.) so the export is useful without cross-referencing.

Spam Detection Management Available in Email Settings

A new Spam Management tab is available under Settings → Email, giving Admins direct control over Gladly's built-in spam detector. If the detector is flagging legitimate Customer emails as spam, Admins can now deactivate it themselves instead of contacting Gladly Support.

When active, the detector identifies potential spam so it can be automatically closed by Rules and bypassed by Gladly AI. Learn more about spam management.

Additional Gladly AI Workflow Outcome Dimensions in Insight Builder

Two new Conversation Metrics dimensions are available in Insight Builder:

• Gladly AI Workflow – Shows which Workflows ran during a Conversation and how each completed, formatted as "WorkflowName: outcome" and comma-separated when multiple Workflows ran.

• Gladly AI Workflow Deployed – A Yes/No flag indicating whether any Workflow ran on the Conversation.

Header Controls Accessible During Incoming Notifications

Agents can now access the top-right header controls (Search, Notifications, Profile Menu) while an incoming Conversation notification is on screen. The Customer name in the notification banner also displays more completely.