Automation Engine

An automation is an if-this-then-that rule that Keme evaluates against every ticket event in real time. When a ticket matches all trigger conditions, the configured actions execute immediately.

Automations replace repetitive manual tasks: tagging tickets, reassigning based on channel, sending acknowledgement replies, escalating to Slack, and more. A well-configured workspace can automate 40–60% of routine agent actions.

Automations are workspace-scoped. They do not run across workspaces. If you need cross-workspace logic, use the webhook action to trigger an external orchestration layer.

Navigate to Automations → New Rule. The rule builder has three sections: Name & Description, Trigger Conditions, and Actions. Each section must be completed before the rule can be saved and activated.

Conditions support all ticket fields (status, priority, channel, tag, assigned agent, SLA state, AI confidence, risk signal level) and logical operators (is, is not, contains, does not contain, greater than, less than).

Rules can have multiple conditions combined with AND / OR logic. Use the group logic toggle to nest condition groups — for example: (Channel is Discord AND Priority is Critical) OR (Tag is VIP).

Automations trigger on ticket events: created, updated, replied (agent), replied (player), status changed, tag added, tag removed, SLA first response breached, SLA resolution breached, and time elapsed.

The time elapsed trigger lets you create scheduled follow-ups — for example, send a check-in reply if a ticket remains in Pending state for more than 48 hours without a player response.

Trigger evaluation is event-driven and real-time. There is no polling delay. However, if multiple conditions depend on the same event (e.g. tag added), the evaluation order follows rule priority, which is configurable in the rule list view.

Available actions include: assign to agent, assign to team, change priority, add tag, remove tag, send reply (from template or custom body), send webhook, create Jira issue, post Slack message, and close ticket.

Multiple actions can be chained in a single rule. Actions execute in order from top to bottom. If an action fails (e.g. the Jira integration is misconfigured), subsequent actions still run and the failure is logged under Automations → Logs.

The send reply action supports template variables: {{ticket.id}}, {{player.name}}, {{workspace.name}}, {{agent.name}}, and {{ticket.url}}. Variables are resolved at execution time.

Before activating a rule, use the Test Rule button to simulate it against a specific ticket. Enter a ticket ID and the rule builder evaluates all conditions against that ticket's current state, showing you which conditions passed and which actions would fire.

Test mode does not execute actions — it is a dry run only. You can test multiple tickets to validate your logic across different scenarios before going live.

Once activated, new rules apply to all subsequent matching events. They do not retroactively apply to historical tickets. If you need to apply a rule to existing tickets, use the Bulk Operations feature with manual filters.

Every automation execution is logged under Automations → Logs. Each log entry records the rule name, the triggering ticket, the timestamp, the conditions that matched, and the actions executed (or their failure reason).

Logs are retained for 30 days. If you need longer retention for audit purposes, configure the webhook action on a logging rule that streams all automation events to your own storage.

The Automations overview page shows each rule's execution count, success rate, and last executed timestamp. Rules with a high failure rate are highlighted in amber — investigate the log details to diagnose misconfiguration.