Conversation priority
The four priority levels
Each conversation carries one priority:
- Urgent. Critical issue blocking the customer. Top of inbox; tight SLA target (e.g., 5-minute first response).
- High. Important but not blocking. Customer waiting actively. Standard SLA.
- Normal. Default. Most conversations.
- Low. FYI or background. Loose SLA.
Priority and channel together drive SLA policies.
Setting priority manually
Three paths:
From the conversation header.
- Open the conversation.
- Click the priority indicator (top-right of the header).
- Pick a level.
Keyboard shortcut. With a conversation focused, press P then 1 (urgent), 2 (high), 3 (normal), or 4 (low).
Via the API.
curl -X PATCH https://api.askvault.co/v1/conversations/conv_xxx \ -H "Authorization: Bearer ak_xxx" \ -H "Content-Type: application/json" \ -d '{"priority": "urgent"}'Changes apply within 30 seconds across the dashboard and SLA tracking.
Auto-priority rules
Configure rules to set priority automatically when conversations match conditions:
- Open Dashboard > Automation Rules > Create Rule.
- Trigger: new conversation created (or message added).
- Conditions (any combination):
- Channel equals X.
- Audience tag equals Y.
- Message contains keyword Z.
- Customer has tag W.
- Action: set priority = urgent / high / normal / low.
Example rules:
- Tag
enterprise_customerand channelemail→ priority = high. - Message contains "down", "outage", "broken" → priority = urgent.
- Audience tag
free→ priority = low. - Customer tag
vip→ priority = urgent regardless of channel.
Rules evaluate in order; first match wins.
What priority affects
Three operational behaviors flow from priority:
- Inbox sorting. Urgent appears at top; low at bottom. Saved Inbox Views can filter by priority.
- SLA targets. SLA policies use priority as a condition (e.g., urgent gets 5-min first-response target; normal gets 1-hour).
- Routing. Automation rules can assign urgent conversations to senior agents only.
- Notifications. Urgent conversations send agent notifications immediately; lower priorities batch.
Sentiment-based priority bumps
The sentiment_router skill detects customer frustration mid-conversation. When sentiment crosses a threshold (default: very negative), the skill can auto-bump priority:
- Customer mentions "cancel my account" or uses strong negative language.
- Sentiment skill detects threshold crossing.
- Skill bumps priority from normal to high (or high to urgent).
- Agent gets notified about the priority change.
Configure under AI Agents > Skills > sentiment_router > Priority Bump.
Priority and skill behavior
Some skills behave differently based on priority:
escalate_to_human. Routes urgent to senior agent queue; normal to standard queue.sentiment_router. Auto-bumps priority on detected frustration.subscription_manager. Auto-priorities billing-failure conversations as high.ticketing_router. Sets Zendesk ticket priority matching AskVault priority.
Bulk priority changes
To bulk-set priority across multiple conversations:
- Filter the inbox (e.g., all conversations tagged
outage). - Select all.
- Bulk Actions > Set Priority.
- Pick a level.
- Confirm.
See bulk operations for the full bulk-action surface.
Priority audit log
Every priority change is logged:
- Author (user, API key, or automation rule).
- Old and new priority.
- Timestamp.
Visible in the conversation timeline and bulk-exportable. Retained 365 days standard, 6 years Enterprise.
Sample workflow
How a typical support team uses priority:
- All conversations default to normal. No work needed.
- Auto-priority rules bump:
- Tag
enterprise_customer→ high. - Message contains "outage", "down" → urgent.
- Audience
free→ low.
- Tag
- Sentiment-based bumps push frustrated customers to high.
- Manual overrides as needed.
- Saved Inbox View "Urgent + High" is the default for the on-shift agent.
- Low-priority backlog processed during slow periods.
About 75 to 85% of priority assignments happen automatically; the rest are manual judgment calls.
Planned features (on the roadmap)
Documented for accuracy:
- Custom priority levels. Today, fixed 4 levels. Custom levels (e.g.,
p0,p1,p2) planned for Enterprise. - Time-based priority escalation. Today, priority is static unless changed. Planned: auto-bump priority if SLA at-risk.
- Customer-set priority. Today, agents and rules set. Planned: customer-facing "how urgent is this?" prompt.
Limits
- Priority levels. 4 fixed values (urgent, high, normal, low).
- Auto-priority rules per workspace. 50.
- Bulk priority change cap. 200 per action.
- Audit log retention. 365 days standard.
Common pitfalls
Every conversation set to urgent. Auto-priority rules too aggressive. Audit rules; only truly urgent patterns should trigger urgent.
Priority doesn't affect routing. SLA policies not configured to use priority as a condition. Check Conversations > SLA Policies.
Customer sees the priority. Priority is internal. If it leaks (e.g., agent mentions in reply), reword. Priority isn't customer-facing.
Rule changes priority retroactively. Rules apply on conversation create or message add. Existing conversations don't re-evaluate unless you trigger a manual re-evaluation.
FAQ
Does the bot use priority for anything?
Not directly. The bot generates answers regardless of priority. Priority drives team routing and SLA tracking, not bot behavior.
Can priority be set by the customer?
Not today. Agents and automation rules set it. Customer-facing priority is on the roadmap.
Does changing priority affect the conversation history?
No. The priority change logs as an audit event but doesn't modify message content.
Can I have more than 4 priority levels?
Today, fixed at 4. Custom levels on Enterprise are on the roadmap.
Does priority sync to my CRM?
Yes if you've enabled CRM integration. The priority field maps to the CRM's matching field (e.g., HubSpot's priority property).