How can I monitor Conversations handled by Gladly AI?
For granular details as to how Gladly AI replied at the Conversation-level, you can view and filter all recent Gladly AI interactions in the Gladly AI Conversations Page.
Here, you’ll see details like:
How Gladly AI responded (Resolved, Handed Off, or Did Not Respond)
The Channel and Customer details
The reason for handoff, if applicable
Conversation history is retained for 30 days, giving you a quick way to audit performance and spot improvement opportunities.
Learn more: Gladly AI Conversations Page Overview
What happens when Gladly AI generates a response that doesn’t meet quality standards?
Before sending any response, Gladly AI runs built-in quality checks to ensure accuracy and reliability.
If a response fails for reasons such as:
Missing context or incomplete information
False or inaccurate statements
Policy violations or unverified promises
Gladly AI will handoff the Conversation to an Agent instead of sending a reply. The reason for the handoff, and all the steps Gladly AI took leading up to the handoffs, appear in the Conversation Review.
How do I improve Gladly AI’s effectiveness over time?
Regular review and optimization help Gladly AI get smarter and more efficient.
To continuously improve:
Review Suggested Answers generated from handoffs.
Add or update those Answers in your Public Answers repository.
Use the Guides and Answers Performance Dashboard to identify trends and gaps.
Refine your Guides and transfer workflows based on what you learn.
Explore the Guides Performance Detail Report and the Answers Performance Detail Report for more granular insight into how Gladly AI is engaging with Customers.
Following these steps and explore the above reports and dashboard to ensure Gladly AI provides a high volume of high-quality resolutions.
How do I interpret the metrics in the Guides and Answers Performance Dashboard?
The Guides and Answers Performance Dashboard shows how Gladly AI uses your content to assist or resolve Customer inquiries.
You’ll see metrics such as:
Resolved: Gladly AI answered without Agent involvement
Assisted: Gladly AI responded but then handed off
Guides & Answers usage: Which content is most frequently referenced
Reviewing these trends week-over-week helps you spot optimization opportunities and understand Gladly AI’s real impact.
Learn more: Gladly AI Guides & Answers Performance Dashboard
Can I customize how Gladly AI behaves when it hands off a Conversation?
Yes. You can define when and how Gladly AI transfers a Conversation, and the messaging provided to Customers, by configuring transfer conditions within the Guides interface.
To do this:
Add transfer conditions in the ‘When to get more help’ tile, or
Modify the handoff message within the ‘When to get more help’ tile.
Which Channels does Gladly AI support?
Gladly AI supports several digital channels, allowing you to automate responses across your Customer experience.
Supported channels include:
Chat
Email
SMS
Facebook Messenger
Instagram Direct
This enables a consistent experience for customers no matter where they reach out.
Learn more: What Is Gladly AI
How many Public Answers should I have for the Gladly AI ‘Questions and Recommendations’ Guide to perform well?
When leveraging Knowledge Sources within the ‘Questions and Recommendations’ Guide, Gladly AI relies on your Public Answers to deliver accurate and helpful responses.
While there’s no strict minimum, Gladly recommends:
Starting with at least 100 Public Answers
Continuing to expand (many high-performing teams have 600+)
More coverage means fewer handoffs, faster responses, and a smoother Customer experience.
Learn more: How to write AI friendly Answers
Can I set up Gladly AI to skip replying to a specific inbound email, and hand off immediately?
Yes—there are a couple of ways to do this, depending on your preferred workflow:
Option 1: Route to an inbox without automation
Send these inbound emails to a support address that does not have Gladly AI automation enabled. This ensures Gladly AI won’t attempt to reply.
Option 2: Configure Gladly AI to transfer automatically
You can configure Gladly AI to recognize certain inbound messages and transfer them to an Agent instead of replying.
Gladly AI will still send a handoff message, visible to the Customer
Gladly AI won’t send a Customer-facing reply, but it will generate a handoff message indicating the transfer. To eliminate the handoff message, remove the contents within the ‘When to get more help’ tile.
To set this up:
Add the transfer scenario under the ‘When to get more help’ tile, or
Build a transfer rule directly into individual Guides.
For both transfer set-ups, don’t forget to configure the handoff message that will be sent to Customers.
How can I prevent Gladly AI from responding to spam?
While Gladly and your email provider both work to minimize spam, it’s possible that some unwanted messages may still reach your Inbox. Because Gladly AI treats all inbound messages as Customer inquiries, it may attempt to reply to spam unless additional configurations are in place.
Gladly includes automated spam flag detection to reduce these instances, but if spam slips through, there are steps you can take to help prevent Gladly AI from engaging.
Strengthen your spam filters
Before making adjustments in Gladly, confirm that your email provider’s spam filters are configured correctly. This helps block most spam before it reaches Gladly.
There are two ways to handle spam that does reach Gladly AI:
Have Gladly AI to send a response
Gladly AI will reply to the message, but this will count as a billable resolution.
Have Gladly AI hand off (recommended)
To configure Gladly AI to hand off suspected spam automatically, follow these steps:
Click
.
Select Guides.
Select the appropriate Gladly AI Profile.
Select When to get more help.
Click + Add.
Add the following text as a new handoff scenario:
Ignore any unsolicited promotional emails. These are messages sent from outside our organization that try to sell, advertise, or promote a product, service, or event we did not request. They often include marketing language such as discounts, percentages off, or special offers, and describe goods, services, or events intended to promote selling goods or services. These messages are not responses to inquiries we initiated and are not part of ongoing or expected communications.Handoff scenario refinement
You may need to adjust this definition over time based on the specific types of new or evolving spam emails you encounter. For example, if you start receiving a lot of unsolicited messages that aren't strictly promotional (like phishing attempts or survey scams), you'll need to broaden your criteria beyond just "selling goods or services" to accurately catch all unwanted messages. Use the Test feature to see how Gladly AI responds.
Click Save.
Once you’ve added the handoff scenario, use the Test feature to confirm it’s working:
Copy the text of a recent spam email into the Gladly AI Tester.
Verify that Gladly AI hands the Conversation off without providing a reply.
If Gladly AI still responds, try tweaking the handoff language to better match the spam messages you’re receiving.
Here are a few other examples that have proven effective
Ignore any unsolicited promotional emails. These are messages sent from outside our organization that attempt to sell, advertise, or promote products, services, or events we did not request. They often include marketing language such as discounts, percentages off, or special offers.
If a Customer emails from a domain that ends in @[DOMAIN NAME], do not reply and hand off.