Skip to main content

AI Spam Filter

AI Spam Filtering for Helpdesk

Automatically filter spam and irrelevant inbound messages before they reach your support team.

Overview

Managing spam and irrelevant conversations can slow down support teams and impact service metrics.

With AI Spam Filtering for Helpdesk, inbound spam messages are automatically detected and separated from genuine customer conversations.

Spam messages are identified using a combination of AI detection and merchant-defined rules. When detected, they are automatically marked as spam and moved to a dedicated Spam view—without triggering agent workflows or unnecessary responses.

This helps teams maintain cleaner queues, faster response handling, and more accurate support reporting.

How It Works?

AI Spam Filtering evaluates only the inbound message received from a customer before running assignment block and assigning to agent.

If the message is identified as spam:

  • A ticket is created if not already created.

  • The ticket is immediately closed and marked as spam

  • No automated or agent reply is sent

  • The ticket appears only in the Spam view

  • Support workflows and queues remain unaffected

This ensures spam stays visible for review while keeping agent workspaces focused on real customer requests.

Supported Channels

AI Spam Filtering currently supports inbound messages from:

  • WhatsApp

  • Email

  • Instagram [DM & Comments]

  • Facebook [DM & Comments]

Key Benefits

Cleaner Helpdesk Operations

Prevent spam and irrelevant messages from cluttering support queues.

Improved Agent Productivity

Agents spend time on genuine customer issues instead of filtering noise manually.

Better SLA & Support Metrics

Spam conversations are excluded from operational reporting, helping improve:

  • First Response Time (FRT)

  • SLA accuracy

  • Queue reporting and performance visibility

Enterprise-Level Control

Merchants maintain full control over how spam is classified and filtered.

Full Transparency

Spam conversations are never deleted and can be reviewed or recovered whenever needed.

Common Use Cases

1. Filtering Promotional or Marketing Noise

Automatically identify and filter:

  • Promotional broadcasts

  • Marketing messages from other brands

  • System-generated notifications

  • Non-customer outreach

2. Blocking Irrelevant Customer Messages

Reduce distractions caused by:

  • Greetings with no support intent

  • Wrong-number messages

  • Abusive or offensive content

  • Non-support-related conversations

Spam Configuration Options

  1. Navigate to Helpdesk Settings.

  2. Click AI Spam Filter.

  3. Click New Spam to create a new spam type.

  4. Select the channels where you want the spam filter to be applicable.

  5. Sample Messages can be added to identify the spam and Non-spam Messages

Channel-Specific Rules

  1. Email - Spam check is performed at the customer's first UIC.

  2. WhatsApp - Spam check is performed when the AI session ends.

  3. Instagram DMs - Spam check is performed when the AI session ends.

  4. Facebook DMs - Spam check is performed at the customer's first UIC.

  5. Comments (FB & IG) - Spam check is performed at the customer's UIC.

For all channels, a spam check is also performed during agent assignment.

Enable the Spam Check Only During Agent Assignment checkbox to perform the spam check only when the ticket is assigned to an agent.

Test Spam Filter

Select the Channel and Enter the test message to check if the message is marked as spam or non-spam message based on the configurations you set.

Manual Spam Marking

Agents can manually mark a ticket as spam by clicking Close as Spam.

  • Agents can click Close as Spam to manually mark a ticket as spam.

  • Tickets from all supported channels can be manually marked as spam.

  • The Chat Activity records how the ticket was marked as spam, indicating whether it was identified by the AI Spam Filter or manually marked by an Agent, along with the agent's name.

Recovering Spam Tickets

Spam tickets remain fully accessible.

If a conversation was incorrectly marked as spam, agents can recover it by:

  1. Click on Spam tickets

  2. Open the ticket

  3. Mark the ticket as not spam

Once marked as not spam, the ticket will be available in All tickets

Guardrails for Spam Filter

The spam check is skipped in the following cases:

  1. Recently closed ticket: If the customer had a ticket that was closed within the last 7 days without being marked as spam, the spam check is skipped.

  2. Recent order: If the customer has placed an order within the last 30 days, the spam check is skipped.

  3. Open ticket: If the customer has any open tickets in the same/another channel, the spam check is skipped.

Frequently Asked Questions (FAQs)

1. What is AI Spam Filtering in Helpdesk?

AI Spam Filtering automatically detects and filters spam or irrelevant inbound messages before they reach support agents. Messages identified as spam are marked accordingly and moved to a dedicated Spam view.

2. How does spam detection work?

Spam detection uses a combination of:

  • AI-based classification

  • Merchant-defined spam rules

The system evaluates the inbound message received from a customer before running assignment block and assigning to agent and determines whether it should be treated as spam.

3. Which channels support AI Spam Filtering?

AI Spam Filtering currently supports:

  • WhatsApp

  • Instagram

  • Email

  • Messenger

Live Chat is not supported in this version.

4. Does spam filtering apply to outgoing messages?

No. Spam filtering applies only to inbound messages.

Outgoing messages and agent replies are not evaluated for spam.

5. When is spam detection triggered?

Spam logic is evaluated on UIC and just before assignment block runs based on the last 5 messages .

6. What happens when a message is marked as spam?

If a message is identified as spam:

  • A ticket is created

  • The ticket is immediately closed and marked as spam

  • No reply is sent

  • No workflows or automations are triggered

  • The ticket appears only in the Spam view

7. Will spam messages affect SLA or First Response Time (FRT)?

No. Since spam tickets are automatically classified and excluded from regular support handling, they do not add unnecessary noise to SLA and FRT metrics.

8. Are spam tickets permanently deleted?

No. Spam tickets remain fully visible in the Spam view and are not deleted.

This ensures teams retain visibility and auditability.

9. Can agents recover a spam ticket?

Yes. If a valid customer conversation is incorrectly marked as spam, agents can recover it by Marking them as not Spam from the Spam view.

Once assigned, the ticket becomes available for normal support handling.

10. Can merchants configure spam rules?

Yes. Merchants can customize spam filtering through:

  • Up to 10 spam types

  • Channel-specific spam rules

This allows businesses to tailor filtering based on their support operations.

11. Will customers know their message was marked as spam?

No automatic response is sent when a message is marked as spam.

The ticket is silently classified and moved to the Spam view.

12. Can merchants create different spam rules for different channels?

Yes. Merchants can define channel-specific spam rules for supported channels such as WhatsApp, Instagram [DM ONLY], and Email.

13. What if AI incorrectly identifies a genuine message as spam?

Teams can review all spam-classified tickets in the Spam view and manually recover any incorrectly classified conversation by assigning the ticket.

This ensures complete control and transparency.

14. Does spam filtering stop ticket creation entirely?

No. A ticket is still created for traceability.

However, the ticket is immediately closed and marked as spam, preventing it from entering normal support workflows.

For further assistance or to raise feature requests related to AI Spam filter, please contact [email protected]

Related Articles
Did this answer your question?