September Releases

September 23, 2026

Control How Gladly Handles Emails from Addresses Used by Multiple Customers

Admins can now control which inbound email addresses are treated as shared senders directly from Gladly.

Some email addresses send messages on behalf of many different Customers – think shipping notifications, automated marketplace emails, or form submissions forwarded from a partner's website. Without this setting, every message from that address collapses onto a single profile, creating a pile of unrelated Conversations.

New ‘Shared Senders’ option in Admin settings – Go to Settings → Email → Shared Senders to view, add, and remove addresses. Once an address is on the list, each new email from it creates its own Customer and Conversation instead of landing on one profile. Replies within an existing Conversation are unaffected, and email already received does not move.
Add from the Customer Profile – Agents with the ‘Channels’ Admin role can add an address to the shared sender list directly from the Customer Profile overflow menu, without leaving the Conversation. A confirmation modal explains the change plainly and lets the Agent pick which address to list when the profile has more than one.

Changes take effect immediately, so your team can keep Customer Profiles clean and Conversations separated as soon as a new shared address appears.

Run Multiple Customer-Idle Workflows in Sequence

The ‘On Customer Idle’ trigger can run more than one Workflow as a Customer stays silent, each after its own delay counted from the Customer's last message. For example, ask whether the Customer is still there after 2 minutes, then send a closing message after 5 minutes. Any reply from the Customer stops the trigger from firing.

Signals: See Every Reviewer's Verdict in One Place

When multiple people review the same Signals result, the panel now shows all of their input instead of just the latest.

Full review history – Each result lists up to five recent reviews with the reviewer's email, their verdict, and any corrected label when they disagreed with AI.
Your reviews stay yours – Your own verdicts show only your decision and never overwrite what another team member recorded.
Results no longer go blank – Previously, heavily reviewed runs could lose their review data entirely. Each result now stores its reviews independently.

Archive Audiences You Are Not Using

Audiences that are not assigned to any help centers, Chat widgets, or Answers can be archived. Archived Audiences no longer appear in dropdown selectors, keeping your lists clean. You can unarchive an Audience at any time to restore it.

Audience active or archive status is also now traceable via API.

Three New Gladly Team Display Languages: Dutch, French, and Polish

Agents can now set their Gladly Team interface language to Dutch, French, or Polish. To update the language display click the Agent avatar → Profile → Language.

The setting is per-Agent and does not affect Customers or other members of your team – menus, buttons, timeline entries, and date formatting all follow the selected language.

Add Topics and Custom Attributes to Tasks at Creation

Agents can add Topics and custom attributes to a Task while creating it, rather than only after the Task has been saved. This means Tasks are categorized from the start, with no extra step to go back and tag them.

Phone Type Support for Custom Channel Integrations

Custom channel integrations can tell Gladly what kind of phone number a Customer's identifier is – mobile, home, or office – by sending phoneType alongside a PHONE_NUMBER profile identifier on an inbound message.

Previously, a phone number supplied this way was stored without a type, which meant a returning Customer could not be matched back to their existing profile – their messages could start a new profile instead of continuing on the existing one, and the number was not usable for SMS. Marking a number as mobile now links repeat messages to the right Customer and, where a matching Customer already exists with that number, attaches it to their existing profile rather than creating a duplicate.

Numbers sent without phoneType behave exactly as they do today.

Updated App Platform Integrations: Ordergroove (v2.3.5)

An update to the Ordergroove integration adds a full subscription and order Card to the Customer Profile, twelve compose-menu forms, and seventeen actions for Gladly AI.

Set up from Apps → Create Configuration, no developer support required.

New view for AI Sessions

We’re introducing an updated view for the AI Sessions page, available in UAT/Sandbox environments, and widely accessible soon.

The new view will give you a more flexible way to explore your AI session data. It will help you:

Understand session activity more easily with key metrics, rates, and trends.
Filter and group sessions to explore your data from different angles.
Quickly access the sessions behind the numbers for further investigation.

When the new view becomes available, the current AI Sessions view will still be available, and you’ll be able to toggle between the two whenever needed.

Email a Third Party from a Task Without Leaving the Conversation

Agents will be able to email a third party – a vendor, a shipper, a partner – directly from a Task without touching the Customer's Conversation.

Threaded on the Task – The outbound email and all replies live on the Task, so the Customer's Conversation stays clean and the exchange is easy to find later.
Replies route back automatically – When the third party responds, the reply lands on the Task and the Task's follow-up date updates, keeping the work visible without the Agent needing to check manually.

Accessible by selecting “Start side conversation” from the compose (+) menu, soon.

Avoid Duplicate Work with Customer Profile Viewer Awareness

Agents will be able to see which other Agents are on the same Customer Profile, so they can avoid sending duplicate replies or stepping on each other's notes.

A pill on the Customer Profile shows the other Agent's avatar, name, and how long they have been there. When two or more are present, stacked avatars and a count appear instead, with a hover list showing each Agent's name, activity – viewing, composing a reply, or adding a note – and duration. The indicator is awareness only; it does not lock the profile or prevent any action.

Stay tuned for more details!

[Early Access] Shopify Product Search for Gladly AI

AI can search your Shopify catalog directly so product questions get answered with up-to-date store data.

Real-time catalog data – Product cards with the real image, current price, working link, and true variant and inventory availability, pulled from Shopify on each request.
Semantic, typo-tolerant search – Customers can ask naturally ("Do you have this in size 8?") and Gladly AI matches them to the right products.
Custom fields included – Details that live only in your custom fields – care instructions, dimensions, ingredient lists – are surfaced to Gladly AI so it can answer questions web search can't.
Web search stays in its lane – Product questions use the Shopify catalog; policies, FAQs, and blog content still come through web search.

Shopify Product Search isn't a separate add-on – it gives Gladly AI a direct, real-time connection to your Shopify catalog for product questions, alongside the web search and Answers it already uses. To test it early, reach out to Gladly Support to get started.

Post-Chat Surveys Send After Automatic Greetings and Quick Replies

Fixed – A post-chat survey would not be sent if the last thing the Customer saw was an automatic greeting or a quick reply rather than a response from the AI Agent. Surveys now go out after any period of Customer silence.

AI Agents and Agent Versions Pages Load Faster

Fixed – The AI Agents and the Agent Versions pages load faster for organizations with large Agent configurations and many saved versions. The deployment banner on both pages was reading each recent version's full configuration just to show its deployment status and save dates, and now reads only those two fields – removing 25 large reads per page load. No change to what appears on either page.

AI Agent and Workflow Saves Are Faster

Fixed – Saving an AI Agent or a Workflow is now noticeably faster, especially for larger configurations. Workflow saves that take unusually long no longer report "Save failed" when the save actually went through – the editor now indicates that the save is taking longer than usual and suggests reloading to check.

Translated Chat Messages Now Toggle Correctly

Fixed – On Chat messages sent with translation, "See message in English" and "See original message" now switch the message body instead of only changing the menu label.

PRODUCT

Signals: Evaluate Gladly AI Session Quality at Scale

Signals is now available to all Gladly AI Customers at no additional charge, subject to fair-use limits. It lets you evaluate AI Sessions at scale against criteria you define—helping you measure performance and uncover feedback, trends, and other insights from real Customer conversations.

Define evaluation criteria in plain English – sentiment, policy adherence, Customer feedback on a product change, or anything else you want to track – then run them against hundreds of Sessions at once.
Pass/fail at a glance – See breakdowns across every evaluation, giving you continuous visibility into how your AI is performing at a scale no team could match manually.
Full reasoning on every result – Click into any evaluation result for the complete transcript and the AI reasoning behind the pass/fail decision.
Monitor and surface insights – Track AI Session trends and uncover Customer insights directly within Gladly.

RESOURCE

See Signals at Work

Want to see Signals in action? This video walks through how to define evaluation criteria, run them against your AI Sessions, and drill into results – so you can start monitoring AI performance at scale.

WATCH ON DEMAND

ARTICLE

Balsam Hill x Gladly: How to Turn Product Questions into Buying Moments

By combining pre-sale shopping assistance and postsale service into a single AI conversation, Balsam Hill boosted its customer resolution rate to 91% and drove six figures in add-to-cart value within 30 days.

REVIEWS

You're In Good Company

Love using Gladly? Share your experience on G2 Crowd to help others discover how Gladly empowers better customer connections.

REVIEWS

Using Gladly with your Shopify store?

Using Gladly with your Shopify store? We'd love your feedback there too! Leave us a review on the Shopify App Store and help other merchants discover how Gladly creates seamless, revenue-driving conversations.

RESOURCE

Gladly Referral Promotion

Love using Gladly? Now’s your chance to share the experience with others in your professional network. Submit your B2C or e-commerce referral to receive up to $2,000! Visit the Refer-A-Friend program to learn how.

September 16, 2026

Signals: Evaluate Gladly AI Session Quality at Scale

Signals is now available to all Gladly AI Customers at no additional charge, subject to fair-use limits. It lets you evaluate AI Sessions at scale against criteria you define – helping you measure performance and uncover feedback, trends, and other insights from real Customer Conversations.

Define evaluation criteria in plain English – sentiment, policy adherence, Customer feedback on a product change, or anything else you want to track – then run them against hundreds of Sessions at once.

Pass/fail at a glance – See breakdowns across every evaluation, giving you continuous visibility into how your AI is performing at a scale no team could match manually.
Full reasoning on every result – Click into any evaluation result for the complete transcript and the AI reasoning behind the pass/fail decision.
Monitor and surface insights – Track AI Session trends and uncover Customer insights directly within Gladly.

Refreshed Workflows List and Editor

Appearance updates make Workflows easier to scan and navigate. These are visual changes only – your Workflows, steps, and data are untouched.

Redesigned Workflows list – The Workflows page is now full-width with search and filters on separate rows, giving long names and descriptions more room. Labels follow sentence case to match the rest of the product.
Editor improvements – Steps inside a Workflow now read like a numbered outline (1, 2, 3, with nested steps numbered 3.1). Step types (Extract, If, Ask, Say) are prominent at the start of each line, variables show their name instead of a technical path, and common actions like duplicate and delete are buttons on the row instead of buried in a menu.

Inspect Summary Explains AI Judge Decisions

When a Workflow ‘If’ step asks an AI judge a plain-language question, the Conversation Review panel now shows a short rationale for the answer next to the condition, instead of leaving that column empty. Anyone reviewing a Conversation can see why a judge matched or did not match without digging further.

Dynamic IVR for Voice AI

Dynamic IVR puts Voice AI at the front of the phone experience so Customers skip long menus and get help immediately. When a human is needed, the call routes to the right Inbox based on Conversation context, and Agents receive a summary at handoff so the Customer does not repeat themselves.

No additional fees – existing assist and resolution charges still apply, and you’re only charged when Voice AI sends an AI-generated message to the Customer. If you’re interested in moving Voice AI to the top of your IVR, reach out to your Account Manager or Gladly Support.

Self-Serve Configuration for Customer Profile

Admins can now add, remove, and reorder the cards that appear on the Customer Profile from Settings → Agent Experience → Customer Profile – no Support ticket needed. A live preview shows the layout as you work, so changes are visible before you save.

Drag-and-drop card ordering – Rearrange cards and save; the new order takes effect immediately across all Customer Profiles. Cards with no data for a given Customer are hidden automatically, so the profile stays clean.
Add and manage custom attribute cards – Create a new card, search for existing custom attributes to include, or register a new attribute by entering its field name and display label. Mark any attribute as Agent-editable so Agents can update values directly from the Customer Profile. An advanced tab lets you turn an attribute into a clickable link or connect it to an App Platform app.
Add App Platform cards – Select an installed app from a dropdown to surface its data – such as order history from an OMS – on the Customer Profile. When a new major version of an app is available, Admins can switch to it here without contacting Support or engineering.
Manage all attributes in one place – A dedicated management view lists every custom attribute on the card, where you can change display labels, toggle Agent editability, and configure advanced options.

Five New Fields for Gladly AI Sessions in Insight Builder

Gladly AI Sessions reporting adds five new fields in Insight Builder explores:

First Guide Used and Last Guide Used – The first and last Guide an AI Session invoked, plus their corresponding IDs. These single-valued fields group cleanly in reports, where the existing comma-separated Guides Used list does not.
Guardrails Used and Last Guardrail Used – Which Guardrails fired during a Session and which one fired last.

Note: The underlying datasets behind Gladly AI Sessions and Session Messages now refresh once daily at 10:00 UTC rather than every four hours. Same-day activity may take longer to appear than it did previously.

Gladly AI Sessions Detail with Messages Report: Simpler Message History Format

The “Message History” field in the Gladly AI Sessions Detail with Messages report now returns plain text instead of JSON. The content is the same – only the format has changed, making it easier to read without parsing. If you have downstream processes that consume this field as JSON, update them to expect text.

New and Updated App Platform Integrations: RepairShopr and Skio

One new integration and one major update bring repair-ticket management and expanded subscription actions into Gladly.

Set up from Apps → Create Configuration, no developer support required.

RepairShopr – View a Customer's repair tickets on the Customer Profile with ticket number, status, subject, problem type, promised date, assigned tech, and location. Agents can start a repair ticket, update a ticket's status, add an internal or Customer-facing note, and book an appointment – all from the Conversation.
Skio (v5.0.0) – A major update turns the read-only Subscriptions card into a working surface. Eleven compose-menu forms now cover the full subscription lifecycle. Gladly AI gains twenty-seven actions including product-line edits, variant swaps, payment-method changes, and commitment offers. Seven merchant settings control every money-moving action – all off by default until an Admin enables them.
- Note: this is a major version; upgrading from a previous Skio version requires choosing the new app name (gladly.com/SKIO) and re-enabling the "Ship now" setting if your team uses it.

To get started, configure an app via Settings → App Developer Tools → Apps.

New view for AI Sessions

We’re introducing an updated view for the AI Sessions page, available soon.

The new view will give you a more flexible way to explore your AI session data. It will help you:

Understand session activity more easily with key metrics, rates, and trends.
Filter and group sessions to explore your data from different angles.
Quickly access the sessions behind the numbers for further investigation.

When the new view becomes available, the current AI Sessions view will still be available, and you’ll be able to toggle between the two whenever needed.

Avoid Duplicate Work with Customer Profile Viewer Awareness

Agents will be able to see which other Agents are on the same Customer Profile, so they can avoid sending duplicate replies or stepping on each other's notes.

Stay tuned for more details!

[Early Access] Shopify Product Search for Gladly AI

AI can search your Shopify catalog directly so product questions get answered with up-to-date store data.

Real-time catalog data – Product cards with the real image, current price, working link, and true variant and inventory availability, pulled from Shopify on each request.
Semantic, typo-tolerant search – Customers can ask naturally ("Do you have this in size 8?") and Gladly AI matches them to the right products.
Custom fields included – Details that live only in your custom fields – care instructions, dimensions, ingredient lists – are surfaced to Gladly AI so it can answer questions web search can't.
Web search stays in its lane – Product questions use the Shopify catalog; policies, FAQs, and blog content still come through web search.

Tasks Created via API Record Topic Associations

Tasks created with Topics through the API (POST /api/v1/tasks and POST /api/v1/customers/:id/tasks) now honor topicIds on create, so Topic associations are recorded for reporting. Previously the field was accepted but silently ignored. The endpoints also now reject unknown, disabled, or category-only Topic IDs with a clear error.

AI Agents Require at Least One Transfer Condition

Fixed – An AI Agent could previously be saved with no transfer conditions, which left it transferring Conversations only when it judged it could not help, with nothing you had written steering that decision. Removing all transfer conditions now prevents saving until you add one back, with a message explaining why.

Workflow Conditions No Longer Treat the Text "false" as True

Fixed – A value of "false" was treated as true in Workflow and Guide conditions. The "has any value" check counted any text as a value, so a variable holding the text "false" would satisfy it, and a branch could run when the author meant the opposite. Text that is empty or only spaces no longer counts as a value either.

AI Judge Steps No Longer Misroute Attachment-Only Messages

Fixed – When a Customer emailed only an attachment with no message text, or wrote through a contact form, the AI could misidentify the message as automated mail. Spam-style routing rules could then apply a spam Topic and close the Conversation, leaving the Customer waiting. AI now recognizes these as real messages from a person.

Conversation Review Panel Shows Formatted Message Text

Fixed – Message text in the Inspect Summary displays formatted, matching what the Customer actually received. Bold text, bulleted and numbered lists, and paragraph breaks previously appeared as raw formatting characters run together on one line.

Customers Get a New Zendesk Session After an Agent Ends Theirs

Fixed – A Customer who kept messaging after an Agent ended their Zendesk session now gets a new Zendesk session instead of their messages going unanswered. Handoff failures that were previously silent are now reported, so the existing error email and the "something went wrong" Chat message reach cases that produced neither before.

If two Sessions hand off within moments of each other, each opens its own Agent ticket and only the newest receives later messages.

Team Assist Image and PDF Analysis Now Works for All Organizations

Fixed – Agents asking Team Assist about a photo or document a Customer sent could get a failure instead of an analysis for some organizations. Image and PDF analysis now works reliably.

PRODUCT

Built-In CSAT Survey Template for Workflows

A new Workflow starter template makes it easy to collect Customer satisfaction scores on AI-resolved Conversations – no external survey integration required.

Ready to use out of the box – The template builds a complete end-of-Conversation survey that asks the Customer to rate the help they received, follows up for an optional comment, and records both for reporting.
Accurate survey reporting – The template includes start and finish marks so reports can show how many Customers were surveyed, how many responded, and how many dropped off.
No manual assembly – Previously, teams had to build each survey step by hand. The template handles the full flow automatically.

Interested in getting started? Reach out to your Account Manager or Gladly Support.

RESOURCE

Build a Workflow in Under 3 Minutes

New to Workflows or want a quick refresher? This short video walks through building a Workflow from scratch – covering the key steps, triggers, and logic so you can start automating Conversations with confidence.

WATCH ON DEMAND

ARTICLE

Balsam Hill x Gladly: How to Turn Product Questions into Buying Moments

REVIEWS

You're In Good Company

Love using Gladly? Share your experience on G2 Crowd to help others discover how Gladly empowers better customer connections.

REVIEWS

Using Gladly with your Shopify store?

RESOURCE

Gladly Referral Promotion

September 9, 2026

Brand Voice Configuration for Team Assist

Admins can configure a brand voice for Customer-facing responses generated through Team Assist. Instead of Agents manually adjusting tone and style on every draft, Team Assist applies your brand's voice automatically – so responses sound like your company from the first suggestion. Less editing for Agents, more consistent Customer interactions, and a faster path from AI draft to sent reply.

Go to Settings → Team Assist → Set up brand voice to get started.

Built-In CSAT Survey Template for Workflows

A new Workflow template makes it easy to collect Customer satisfaction scores on AI-resolved Conversations – no external survey integration required.

Ready to use out of the box – The template builds a complete end-of-Conversation survey that asks the Customer to rate the help they received, follows up for an optional comment, and records both for reporting.
Accurate survey reporting – The template includes start and finish marks so reports can show how many Customers were surveyed, how many responded, and how many dropped off.
No manual assembly – Previously, teams had to build each survey step by hand. The template handles the full flow automatically.

Interested in getting started? Reach out to your Account Manager or Gladly Support.

Conversation Review Panel: Idle Workflow Steps, Action Run Links, and Readable Knowledge Cards

Three updates make it easier to review what happened during an AI Session.

Idle-triggered Workflow steps – Steps from a Workflow that runs after a Customer goes quiet – for example, in the case of a satisfaction survey – now appear in the Inspect Summary. Previously only the part that ran after the Customer replied was visible, so the Workflow looked like it started partway through.
Action run log links – A Gladly Action that an AI Agent called now links straight to that Action's run log, making it easier to debug.
Knowledge searches as readable cards – Knowledge searches now show the Answers and website results they returned as readable cards instead of raw data, so it is clear what the AI Agent actually drew on.

Keyboard Accessibility for Admin Modals

Admin modals across all Gladly Team Admin pages now close on Escape, keep keyboard focus inside the modal while open, and return focus to the control that opened them.

Task Topics Available in Insight Builder

Task Metrics in Insight Builder now include Topics, Topics with Hierarchy, and Freeform Topics, matching the Topic reporting already available for Conversations, along with counts of how many were added, removed, and are currently applied. Freeform Topics on Tasks will populate as Agents begin applying Task attributes.

Team Assist Agent Usage Report

A new report shows how each Agent is using Team Assist, broken down by date – viewable daily, weekly, monthly, or quarterly – and by whether the Agent was actively handling a contact at the time versus ad-hoc use such as looking something up on a Customer Profile. For each Agent, the report tracks:

Team Assist Customers – Distinct Customers the Agent used Team Assist on
Total Team Assist Questions – Questions the Agent asked Team Assist (auto-generated follow-ups excluded)
Suggested Replies – Customer-facing replies Team Assist drafted for the Agent
Informational Responses – Answers Team Assist wrote back to the Agent
Proposed Actions – External actions Team Assist proposed for the Agent to approve
Errors – Team Assist responses that failed

Every response is attributed to the Agent who asked, and a companion Usage Over Time view trends Customers and questions across the selected period. Available as an on-screen dashboard and CSV export.

Access via Reports → Agents → Team Assist Agent Usage.

Add Topics and Custom Attributes to Tasks

Agents can add and remove Topics and custom attributes on a Task directly from the Conversation Timeline. Every change is recorded in the Conversation history for a clear audit trail.

New and Updated App Platform Integrations: AfterShip, Chargebee, and Purple Dot

Two new integrations and one update bring shipment tracking, returns, warranty claims, pre-order management, and subscription billing improvements into Gladly.

Set up from Apps → Create Configuration, no developer support required.

AfterShip (v2.0.0) – A rebuilt integration on AfterShip's public APIs, any merchant can now set it up with their own API key.
Four Profile Cards – Shipments (carrier status, checkpoints, delivery ETA), Returns (approval state, label links), Warranty Claims (resolution progress), and Orders (line items, financials).
Ten compose-menu actions – Approve, reject, and resolve returns; generate prepaid labels; approve and reject warranty claims; predict delivery dates; restart stalled tracking; and more.
Safety switches – Money-moving actions (return rejection, claim rejection, claim processing) are off by default until an Admin turns them on.
Purple Dot – View a Customer's pre-orders on the Customer Profile with status, line items, expected ship dates, shipping address, order totals, and refund history. Agents can issue refunds against exported Shopify orders and share a self-service management link so Customers can change or cancel their own pre-orders.
Chargebee (v1.1.0) – Agents can now choose whether a subscription cancellation takes effect immediately or at the end of the paid term (end-of-term remains the default). The Profile Card now shows which outcome actually occurred and displays trial cancellations.

To get started, configure an app via Settings → App Developer Tools → Apps.

Custom Actions Can Read Contact ID and Chat Onboarding Email

Two new system variables give Custom Actions direct access to data that previously required a workaround.

$contactId – The ID of the current contact, available to any Custom Action without an extra lookup.
$chatOnboardingEmail – The email the Customer entered in the chat onboarding form. Only present when the form collected one – for a Customer identified another way, such as by phone, the variable is left out of the action call rather than sent as an empty value.

Previously these values were only available to Workflows, at times requiring that extra Custom Actions were built to look up information Gladly already had.

Self-Serve Configuration for Customer Profile Cards in Gladly Team, Now Available in UAT

An additional option in Admin Settings is coming to Gladly Team in the next few weeks, giving Admins direct control over a part of the Agent experience that previously required Gladly Support assistance.

Customer Profile card management – Add, remove, and reorder the cards that appear on the Customer Profile from Settings → Agent Experience → Customer Profile.
Changes take effect instantly across all Customer Profiles and only display cards on a profile when the Customer has a value for them.

App Platform cards can also be added from this page, removing another setup that previously required Support.

Signals: Evaluate AI Session Quality at Scale

Later this month, Signals will become available to all Gladly AI customers.

Define what matters – Write evaluation criteria in plain English around the things most important to your operation: sentiment, policy adherence, Customer feedback on a new product change, or anything else you want to track.
Run at scale – Evaluate hundreds of Sessions at once and see pass/fail breakdowns, giving you continuous visibility into how your AI is performing at a scale no team could match manually.
Drill into any result – Click into any evaluation result for the full transcript and AI reasoning behind the pass/fail decision.
Monitor and uncover insights – Track AI Sessions and surface Customer insights, directly within Gladly.

Signals is included with Gladly AI at no additional charge, subject to fair use limits.

[Early Access] Shopify Product Search for Gladly AI

AI can search your Shopify catalog directly so product questions get answered with up-to-date store data.

Real-time catalog data – Product cards with the real image, current price, working link, and true variant and inventory availability, pulled from Shopify on each request.
Semantic, typo-tolerant search – Customers can ask naturally ("Do you have this in size 8?") and Gladly AI matches them to the right products.
Custom fields included – Details that live only in your custom fields – care instructions, dimensions, ingredient lists – are surfaced to Gladly AI so it can answer questions web search can't.
Web search stays in its lane – Product questions use the Shopify catalog; policies, FAQs, and blog content still come through web search.

Copied Guides Route to the Correct Sections

Fixed – Copying a Guide to another AI Agent profile could leave a section's Workflow rules pointing at the original profile's sections instead of the copy's, so the copied Guide's routing did not work.

‘Stop’ Step Target Stays Visible While Editing

Fixed – When editing a ‘Stop’ step's "Go to Agent/Guide" target in the Workflow editor, the currently configured Guide or Agent now stays visible while you search for a replacement, and the menu opens on that option.

Idle and Before-Handoff Workflows No Longer Overwrite Each Other

Fixed – When an AI Agent ran both an idle-time Workflow and a before-handoff Workflow in the same Session, whichever one finished second would erase the other's progress. This could silently discard a satisfaction survey's answers or drop a Customer's request for a human. Both now proceed independently.

Workflows with Action Adapters Returning Lists, Dates, or Times Are Saved

Fixed – A Workflow using an Action Adapter that returns a list, a date, or a time could not be saved. Error messages are also clearer: when a step's output definition is invalid, the editor names the step and the specific field instead of showing an empty message.

"Hand Off to a Specific Inbox" Option Appears in the Workflow Editor

Fixed – The Inbox picker on a "Handoff to human" ‘Stop’ step was not visible within the editor. The picker now appears as intended.

AI Summary Available on Helpdesk Transfer

Fixed – An Agent picking up a transferred Conversation in an external helpdesk could see an empty AI summary instead of the AI's recap of what the Customer said. The recap is now written before the helpdesk is notified, so the Agent starts with the context the AI already gathered.

Signals Criteria Descriptions Now Accept Up to 8,000 Characters

Fixed – Evaluation criteria descriptions now accept up to 8,000 characters (up from approximately 2,000), with a live character counter and a clear inline message when the limit is exceeded.

Single Sign-On No Longer Returns a 404 or Redirects After Being Disabled

Fixed – Single sign-on could return a "404 page not found" error instead of redirecting Agents to their identity provider when more than one organization was configured with the same Gladly URL. Also fixed an issue where Agents were still sent to their identity provider after single sign-on had been turned off for their organization.

Signals Calibration Corrections No Longer Discarded on Verdict Click

Fixed – Clicking a thumbs up/down on a previously corrected calibration result could quietly discard the corrected label and note. Corrections now persist through subsequent verdict clicks and are unaffected by a teammate reviewing the same Conversation.

PRODUCT

Knowledge Sources: One Place to Manage Everything Gladly AI and Team Assist Knows

A new Knowledge Sources page is available directly from the main navigation, giving Admins a single destination to view and manage the knowledge that both Gladly AI and Team Assist draw from when answering Customer questions or supporting Agents.

Previously, connecting Gladly Answers and websites happened across different areas of the platform. The new page unifies this experience, making it easier to see which sources are active and identify knowledge gaps.

RESOURCE

Build a Workflow in Under 3 Minutes

WATCH ON DEMAND

ARTICLE

Forrester: Great Postsale Experiences Multiply Value For Customers And Organizations

Great postsale experiences are a proven revenue multiplier. AI is the engine powering them, giving ecommerce CX leaders a compelling case for smarter service investment.

REVIEWS

You're In Good Company

Love using Gladly? Share your experience on G2 Crowd to help others discover how Gladly empowers better customer connections.

REVIEWS

Using Gladly with your Shopify store?

RESOURCE

Gladly Referral Promotion

September 2, 2026

Knowledge Sources: One Place to Manage Everything Gladly AI and Team Assist Knows

Smarter Conditions and Better Previews in the Workflows Editor

Three updates make Workflows easier to build and review.

Variable path warnings in ‘If’ conditions – Typing a variable path like $.builtin.customerProfile.email into an ‘If’ step’s condition field now raises a warning explaining that the condition compares against that literal text, not the value the path points to. Previously the condition would silently never match.
Compare against another variable – ‘If’ step conditions can compare against another variable instead of only a typed-in value – for example, a purchase date extracted from a receipt against the purchase date on the Customer Profile. Switch a condition's value from ‘Static’ to ‘Variable’ to choose from the same variables the left side offers, on text, number, and date conditions.
Note preview in the tester – Testing a Workflow that adds a Note shows the Note exactly as it will display on the Conversation Timeline – variables filled in and line breaks intact – so you can check the wording before the Workflow goes live. Previously the tester showed only what went into the step, not what came out. Notes added by a Workflow are also now visible on the step's card when reviewing a past Conversation.

Scope Guides to Specific Channels

Guide authors can now limit each Guide to a specific set of Channels. From the Guide editor, choose which Channels a Guide runs on using the drop-down menu next to each Guide's name. Guides left unscoped continue to run on every Channel.

Linked Guide and Section Names in the Conversation Review Panel

Guide and section names in the Inspect Summary are now clickable links. Each one opens the Guide as it was configured when the AI Agent ran it.

Signals Criteria Version History

Signal run detail shows which version of each criterion a run evaluated against, and the Conversation Review panel shows the version each verdict was graded under. Previously, a run displayed whatever the criterion looks like today, so editing a criterion could change what a finished run appeared to have measured.

Three New Exports in the API Export Tool

The API Export Tool (Support Portal) adds three new CSV exports that make it easier to audit and maintain your Gladly configuration.

Export Scheduled Reports – Download a CSV of all scheduled reports in your organization, including who created each report, what it covers, and its configuration details. Useful for Admins doing routine housekeeping or tracking down who owns a specific scheduled report.
Export People Match – Download your existing People Match configuration as a CSV for a complete snapshot of your current matching rules. Useful for auditing, documentation, or planning configuration changes.
Export Entry Points – Download your Entry Point routing setup as a CSV for a complete view of how Entry Points are configured. Useful for auditing which Agent versions are served and how traffic is routed.

Access both exports at support-tools.gladly.tools.

Reports Menu Uses ‘Gladly AI’ Instead of ‘Sidekick’

Reports that previously used "Sidekick" in the reporting menu now say "Gladly AI," completing the product naming change across the platform. This is a labeling change only – no report data, columns, filters, or API identifiers have changed.

Six report titles are updated:
- Conversation Export with Sidekick → Conversation Export with Gladly AI
- Sidekick Answer Performance Detail → Gladly AI Answer Performance Detail
- Sidekick Conversation Time Dashboard → Gladly AI Conversation Time Dashboard
- Sidekick Dashboard → Gladly AI Dashboard
- Sidekick Guides and Answers Performance Dashboard → Gladly AI Guides and Answers Performance Dashboard
- Sidekick Guides Performance Detail → Gladly AI Guides Performance Detail
No breaking changes – Same data, same filters, same columns. Saved links, bookmarks, and reporting API integrations continue to work because the underlying report identifiers are unchanged.

Please note that help center articles may still reference the former Sidekick names until those articles are updated.

Resizable Side Panel in Gladly Team

Agents can now drag the side panel that holds Answers, Team Assist, and Tasks to make it wider. The width is remembered across Sessions and devices, so Agents only need to set it once.

New and Updated App Platform Integrations: Chargebee, Fillout, ReturnGo

Three new App Platform integrations and four updates expand what Agents and Gladly AI can resolve without switching tools – from subscription billing and form submissions to order cancellations and address corrections.

Set up from Apps → Create Configuration, no developer support required.

Chargebee – View a Customer's subscriptions, payment method, invoices, and credit balances from a dedicated Card on the Customer Profile. Agents can send payment-update links, retention offers, upgrade cost previews, and upgrade checkout links, update billing or shipping addresses, and cancel subscriptions – all from the Conversation.
Fillout – An AI-only integration that lets Guides and Workflows look up a Customer's Fillout form submissions by verified email or submission ID, retrieve specific answers, and confirm the submitter's identity – no submission ID needed from the Customer. Admins can restrict which forms are readable.
ReturnGO – Surface a Customer's ReturnGO returns (RMAs) on the Customer Profile with status, items, fees, shipment tracking, refund details, and the full return timeline. Agents can approve, reject, cancel, receive, validate, and complete a return and add notes to its timeline directly from the Conversation.
ShipHero (v1.2.0) – Two new actions let Agents and Gladly AI cancel an order and update a shipping address directly in ShipHero during a Conversation. Lookup actions also now return recipient name and company on shipping and billing addresses.
Rise.ai (v1.1.0) – The Customer Profile Card now displays for Customers with a $0 wallet balance.
Gladly APIs (v1.6.0) – Two new actions: agents returns the list of Agents with ID, name, and email; update_customer patches a Customer Profile's name, image, address, emails, and phones.
appcfg CLI (v5.4.1) – Added support for private OAuth authorization_code grants.

To get started, configure an app via Settings → App Developer Tools → Apps.

Scheduled Reporting Maintenance

We will be performing a reporting infrastructure enhancement on Wednesday, September 2 at 10 PM PT.

During this time, Gladly Reporting features, including OOTB Reports and Insight Builder, will be inaccessible. The core Gladly Team platform and Gladly AI are not expected to be affected.

CSAT Survey Template for Workflows

A new Workflows starter template will make it easy to collect Customer satisfaction scores on AI-resolved Conversations – no external survey integration required.

Ready to use out of the box – The template will build a complete end-of-Conversation survey that asks the Customer to rate the help they received, follows up for an optional comment, and records both for reporting.
Accurate survey reporting – The template includes start and finish marks so reports can show how many Customers were surveyed, how many responded, and how many dropped off.
No manual assembly – Previously, teams had to build each survey step by hand. The template handles the full flow automatically.

Interested in learning more? Reach out to your Account Manager or Gladly Support.

Brand Voice Configuration for Team Assist

Admins will soon be able to configure a brand voice for Customer-facing responses generated through Team Assist. Instead of Agents manually adjusting tone and style on every draft, Team Assist will apply your brand's voice automatically – so responses sound like your company from the first suggestion. This means less editing for Agents, more consistent Customer interactions, and a faster path from AI draft to sent reply.

Self-Serve Configuration for Customer Profile Cards in Gladly Team

Customer Profile card management – Add, remove, and reorder the cards that appear on the Customer Profile from Settings → Agent Experience → Customer Profile.
Changes take effect instantly across all Customer Profiles and only display cards on a profile when the Customer has a value for them.

App Platform cards can also be added from this page, removing another setup that previously required Support.

Signals: Evaluate AI Session Quality at Scale

Later this month, Signals will become available to all Gladly AI customers.

Define what matters – Write evaluation criteria in plain English around the things most important to your operation: sentiment, policy adherence, Customer feedback on a new product change, or anything else you want to track.
Run at scale – Evaluate hundreds of Sessions at once and see pass/fail breakdowns, giving you continuous visibility into how your AI is performing at a scale no team could match manually.
Drill into any result – Click into any evaluation result for the full transcript and AI reasoning behind the pass/fail decision.
Monitor and uncover insights – Track AI Sessions and surface Customer insights, directly within Gladly.

Signals is included with Gladly AI at no additional charge, subject to fair use limits.

[Early Access] Shopify Product Search for Gladly AI

AI can search your Shopify catalog directly so product questions get answered with up-to-date store data.

Real-time catalog data – Product cards with the real image, current price, working link, and true variant and inventory availability, pulled from Shopify on each request.
Semantic, typo-tolerant search – Customers can ask naturally ("Do you have this in size 8?") and Gladly AI matches them to the right products.
Custom fields included – Details that live only in your custom fields – care instructions, dimensions, ingredient lists – are surfaced to Gladly AI so it can answer questions web search can't.
Web search stays in its lane – Product questions use the Shopify catalog; policies, FAQs, and blog content still come through web search.

Workflow ‘Stop’ Step Routes to Guides Across AI Agents

Fixed – A Workflow Stop step that routed to a Guide in a different AI Agent could fail with "Unknown guide" error unless that AI Agent was already published and connected to a Channel. ‘Stop’ steps now route to the selected Guide as long as the Guide is active, regardless of whether its Agent has been published.

Task Audit Items Grouped Under ‘Status & Audit’ Filter

Fixed – Task audit entries are now controlled by the ‘Status & Audit’ filter instead of the ‘Tasks’ filter.

Topics Findable by Name Under Long Category Names

Fixed – A Topic could not be found by typing its name in Search if it sat under a category with a long name. Agents searching the Topic filter would see no results – or see archived Topics instead of the active one – even though the Topic existed and was active. Topics are now found by their own name regardless of the category length.

Task Pill Stays Visible When Switching to Team Assist

Fixed – The Task shortcut disappeared from a Conversation after an Agent switched to the Team Assist panel, leaving no way back to the Task they were working on. Also fixed an issue where clicking a second Task while the Answers panel was open showed the previously selected Task instead of the one clicked.

Answer URLs Clickable in Sent Emails

Fixed – Links in Answers are now clickable when sent on email. When an Agent inserts an Answer containing a URL into an email – including URLs authored in an Answer's Chat/Messaging content and URLs with fill-in values – the URL is highlighted as the Agent composes and delivered to the Customer as a clickable link.

Agent Activity Recorded on Admin Pages

Fixed – Time Agents spent on several admin pages was not recorded. These visits were missing from Agent activity history and utilization, and an Agent's status on the Agents Liveboard did not reflect that they were on the page.

User Role Summary Report Date Reflects Selection

Fixed – Changing the User Role Summary report's date range had no effect; the report always showed the last 3 months. It now returns the months you select. The report also has a new Month Complete column so it is clear when a month is still in progress and its counts are not yet final.

Guides and Answers Performance Reports No Longer Miss the Most Recent Day

Fixed – The Guides Performance Detail and Answers Performance Detail reports could omit the most recent complete day of data. The reports now refresh as each batch of data lands. Gladly AI Sessions report, which had the same issue, now refreshes hourly as well.

PRODUCT

New Reports: User Role Summary + Team Assist Feedback and Usage

Missed last week's release? Two new reports are now available for Admins looking to understand how their team is using Gladly.

Team Assist Feedback and Usage Report (Reports → Gladly AI) – See how Agents are using Team Assist and how helpful they're finding it, including thumbs-up and thumbs-down feedback and the reasons behind negative ratings.
User Role Summary Report (Reports → Agents) – A monthly breakdown of how each User's activity classified them – Support User, Task User, or Information User – so you can see how your team is using Gladly at a glance.

Learn about User Role Summary | Explore Team Assist Feedback and Usage

RESOURCE

On-Demand: Gladly AI Workflows Classroom Session

Weren't able to join us live? The Gladly AI Workflows: Build with Confidence session is now available on demand. It covers everything from core building blocks to advanced logic – including key use cases, testing and troubleshooting tips, and how Workflows fit into your broader AI setup.

Whether you're building your first Workflow or refining what's already live, this is a great way to get up to speed.

WATCH THE RECORDING

ARTICLE

Forrester: Great Postsale Experiences Multiply Value For Customers And Organizations

Great postsale experiences are a proven revenue multiplier. AI is the engine powering them, giving ecommerce CX leaders a compelling case for smarter service investment.

REVIEWS

You're In Good Company

Love using Gladly? Share your experience on G2 Crowd to help others discover how Gladly empowers better customer connections.

REVIEWS

Using Gladly with your Shopify store?

RESOURCE

Gladly Referral Promotion