FAQs: Automatic spam detection in Concierge

Concierge has built a spam detection system to assess whether subscriber responses intend to start a conversation with your brand.

When a subscriber responds to your brand, Attentive assesses whether or not the message is spam. If it is, the message doesn’t route to Concierge agents. This ensures that Concierge agents only respond to subscribers who need help.

How does Attentive determine if a subscriber’s message is spam?

Attentive uses data from a machine learning model and other Concierge spam messages to determine whether a subscriber’s message is spam. Attentive detects spam on the first subscriber response, which would otherwise initiate a conversation.

What are some examples of messages that are detected as spam?

Most spam messages fall into a few categories:

  • Automatic responses from subscribers: This can include messages such as “I’m driving with Focus turned on. I’ll see your message when I get where I’m going.”
  • Subscriber reactions to campaigns and triggered messages: Examples include messages in which the subscriber “loved” or “emphasized” the message.
  • Nonsense messages: Subscribers may accidentally respond with a series of characters or letters that don’t translate into words.
  • Logical messages that were  meant for someone else: If a message is clearly unrelated to the campaign or triggered message (such as a grocery list) Attentive detects these as spam.

What happens if a subscriber sends a valid question before/after a spam message?

Concierge agents are routed messages any time we detect it is not spam. When a subscriber responds with a message that is not detected as spam, the message is routed to a Concierge agent regardless of spam sent before/after that message.


Articles in this section

Was this article helpful?
0 out of 0 found this helpful