Control when Fin hands off conversations to your team by defining deterministic Escalation Rules using attributes and customer data, or by creating natural language Escalation Guidance for more flexible scenarios.
This helps you provide timely support when customers have high-intent queries, express frustration, or meet specific data criteria you define.
Escalation methods
There are two methods: Escalation Rules and Escalation Guidance, each designed for different use cases.
Escalation Rules
Escalation Rules trigger a handoff to your team when defined data attributes are detected in a conversation. It’s perfect for situations where you have structured data about your customers or the conversation itself.
For example, you could set up a rule to automatically escalate a conversation if:
The detected value for the Fin attribute "Sentiment" is "Negative".
The detected value for the Fin attribute "Issue Type" is "Bug".
The conversation data contains an order total over a certain value.
A custom data attribute like "VIP_customer" is true.
Tip: You can view and manage Fin attributes, like "Sentiment" and "Topic", by navigating to Train > Attributes.
To create an Escalation Rule:
Navigate to the Train > Escalations.
Click + New under Escalation Rules, add a condition to define the triggers for the escalation. These conditions can be based on Users, Company, or other conversation data.
Once you’ve configured your conditions, click Save and Enable.
Pro tip: To safely test a new rule before it affects all customers:
Add a condition to the rule that targets only your team (e.g.,
Email contains "@yourcompany.com").Click Save and Enable.
Go to the Preview window on the right and set the "Testing as" user to an email that matches your condition (e.g.,
your-email@yourcompany.com).Type a message that should trigger the rule and check the Event log tab. You'll see a confirmation that your rule was correctly followed.
Once you've confirmed it works, you can remove the email condition to activate the rule for all customers.
Escalation Guidance
This method triggers an escalation using natural language scenarios or customer behaviors. This is ideal for responding to a customer's intent or actions when specific data attributes aren't available.
For example, you could set up a scenario based rule to automatically escalate a conversation if a customer:
Uses words that indicate frustration, like “angry” or “not working”.
Asks to speak directly with a human teammate.
Repeatedly visits your pricing or cancellation page.
To create Escalation Guidance:
Navigate to Train > Escalations.
Click + New under Escalation Guidance, add fine-tune Fin’s escalation behaviour in specific scenarios using natural language.
Once you’ve written your guidance, click Save and Enable.
Tip: For advice on writing clear and effective guidance, check out our Fin Guidance best practices article.
Escalation Guidance examples
Moments when Fin should route customers straight to your team.
Be extra sensitive to customer frustration - if the customer expresses irritation, disappointment, or uses words like 'ridiculous' or 'waste of time' escalate immediately.
If a customer asks to speak with a human, offer to connect them to a support agent rather than escalating immediately.
Offer to route the customer to a human agent if they ask the same question twice.
If the customer mentions a competitor's name, route to a human agent and let them know you're connecting them with a team member who can better assist.
Using Escalation in workflows
To make it easy for you to configure routing and handover behavior when Fin escalates a conversation to the team, you can create branches in workflows based on which Escalation has been applied. This gives you full control of how Fin routes to the team for specific scenarios.
Simply create a branch in your workflow for each type of escalation, and combine it with other filters like customer type or language for even more control.
To set this up:
Create your escalation guidance in Train > Escalation.
Then go to Deploy, open your Fin workflow and add Branches after the Let Fin answer step.
Choose “Escalation guidance applied” as your condition.
Select the escalation guidance you want to branch on (e.g., “Refund requests”).
Assign each branch to the right action in your workflow.
Differentiating between Escalation and Workflows
While Escalation Rules and Guidance handle the detection and initiation of escalations, your Fin workflow is used to manage the steps that follow the escalation.
For example:
Escalation Rule/Guidance: Specify when and why escalation should occur.
Workflow: Handle post-escalation actions such as collecting additional information from the customer, creating a ticket, or routing the issue to the appropriate team.
This separation ensures that the escalation process is both efficient and customizable for your needs.