Inbox
Reply Classification
How Obert's AI classifies inbound LinkedIn replies and how to use classifications.
When a lead replies to a campaign message, Obert's AI analyzes the response and classifies it automatically. This helps you prioritize which conversations need attention.
Classification types
| Classification | Meaning | Example |
|---|---|---|
| Interested | The lead expressed interest or asked for more info. | "That sounds interesting, can you tell me more?" |
| Not Interested | The lead declined or asked to stop. | "Thanks but we're not looking for this right now." |
| Follow-up Needed | The reply is ambiguous or requires judgment. | "Can you reach out again in Q3?" |
Where classifications appear
- Inbox — classification badges appear on conversations in the chat list.
- Campaign leads — the lead's campaign status shows whether they replied with interest.
- Lead detail — the activity timeline includes the classification.
How classification works
- A lead replies to a campaign message.
- Obert's AI reads the reply content.
- A classification is assigned with a confidence score.
- The conversation is tagged with the classification.
Classification happens automatically within seconds of receiving the reply.
Override a classification
If the AI got it wrong:
- Open the conversation in the inbox.
- Click the classification badge.
- Select the correct classification.
Your override takes precedence. The original AI classification is preserved for reference.
Use classifications to prioritize
- Filter inbox by classification — click Filter and select "Interested" to see only hot replies first.
- Daily review workflow — start with "Interested" replies, then "Follow-up Needed", then handle "Not Interested" last.
- Campaign analytics — classification data feeds into campaign metrics (reply rate, positive reply rate).
System labels
Classifications are also applied as system labels on conversations:
- Interested (green)
- Not Interested (red)
- Follow Up (yellow)
These labels work the same as custom labels — you can filter by them in the inbox.