Manual assignment is fine until the team grows. With two operators, someone sees a conversation and takes it. With ten operators, the same habit turns into a queue nobody quite owns. Urgent conversations sit behind easy ones. Slack gets noisy. People assume somebody else replied. The product has to make ownership obvious.
A notification is not ownership
A Slack message tells the team that something happened: a new conversation, a reply, an assignment, an SLA risk, or an AI handoff. It does not decide who owns the work. Convor keeps those jobs separate. Slack gives visibility. The routing layer decides where the conversation should go.
The inputs that matter
- Department: sales, billing, onboarding, and technical support need different queues.
- Presence: assigning work to an offline operator hides the delay.
- Workload: round-robin is fair only when the load is similar.
- Priority: a serious escalation should not wait behind a routine question.
Those inputs change during the conversation. An operator goes offline. A visitor sends another message. A manual takeover happens. The first-response target starts to look risky. That is why assignment needs history. The team should be able to see who owns the conversation now and how it got there.
SLA should trigger action
SLA is not useful if it only becomes a red number after the customer is already annoyed. A good policy warns early, posts to the right channel, and can reassign the conversation when the current owner is unavailable. In Convor, SLA state sits next to routing and auto-reassign logic because the fix for a breach is usually operational. Add coverage. Move the conversation. Adjust the rule.
Rollout order
Start with visibility. Send Slack notifications for new conversations and SLA risk. Then add departments and basic assignment rules. Only after that enable auto-reassign for narrow cases, such as offline owner or missed first response. Keep manual takeover available. Automation should remove obvious waiting time, not stop experienced people from using judgment.
Routing works best when people can predict it. If operators understand why a conversation landed with them, they trust the system. If assignments change without a visible reason, they work around it. The boring version is the reliable one.
Get new posts in your inbox
No spam. Unsubscribe anytime.
