Routers allow you to automatically distribute interactions for QA reviews or AI processing. In this guide, you’ll learn how to create, configure, and manage routers so your quality workflows run smoothly and efficiently.
🛠️ Setting Up a Router
Step 1 — Basics
a) Name Your Router
Choose a clear, descriptive name.
Examples:
Weekly QA – Sales Team
Daily Compliance Review – Banking
b) Filters
Examples: Phone Calls Only, VIP Customers, Escalations
c) Choose Distribution Strategy
In this Routers Overview & Concepts article, we describe the differences between each distribution strategy. (Only for Manual QA - Scorecards ⚠️ )
Options:
Round-Robin
Random
Sequential
d) Type (Products)
Options:
Transcription
AI Process
AI metrics
AI Tagging
Scorecard
Step 2 — Assignments
a) Choose which Teams or agents will be sampled.
You can select the entire team and choose how many interactions you’d like to evaluate for each of the agents who belong to that team.
If you select a team, you can also choose specific agents below in case you’d like to run a higher number of evaluations for certain individuals. For example, you might want 1 interaction per week for all team members, but 3 per week for Mike since he’s new.
Example:
Agent | Sampling |
|---|---|
Sarah Johnson | 1 interactions |
Mike Chen | 3 interactions |
Maria Garcia | 1 interactions |
b) Assign Reviewers (Scorecard Routers)
Who will receive the interactions to evaluate, and what percentage of the load each auditor gets.
c) Allocation
Total must be 100%
Example:
Reviewer A → 60%
Reviewer B → 40%
Step 3 — Schedule
a) Time Window (days)
Days back from the current date from which interactions are queried.
A good best practice is to set a look-back window of at least 7 days or more, so Oversai has a better chance of finding tickets in the source of truth that match the characteristics the router is searching for.
b) Set Expected Completion Deadline (days)
Number of days within which auditors are expected to complete the assigned interactions.
c) Set Schedule (Execution days)
Here you can define the days when you want the routers to run, as well as configure the execution time and timezone.
For example:
You may want the router to run every Monday at 8:00 AM Eastern Time.
d) Activate
Set router status to Active.
The router will run automatically on the configured schedule.
Assignment Forecast
In the Assignment Forecast button, you can see a calculation based on:
Expected interactions × active days per week × 4 weeks
This provides an estimate of how many interactions will run during the month based on your configured routers.
You will also be able to see the hard limit for the AI volumes you have contracted.
Note: This is an estimate and may vary depending on the volume of interactions, filters, and router configurations.
Recommendations
To ensure successful router execution:
Sufficient Interactions
Make sure your filters return enough interactions. Otherwise, the router may not be able to assign the desired number.Relevant Filters
Verify that the filters include only the interactions you want—ideally with status closed or resolved, so the AI has enough data to evaluate.Expand time window
To find more potential tickets to analyze.
Previously Processed Interactions
Interactions already processed will not be included again for the same AI product.Example: An interaction processed by AI Tagging cannot be reprocessed by another AI Tagging router.
Exception: AI Process routers can reprocess interactions that were handled by a different AI Process.
Monitoring Router Execution
Track the progress of router assignments in the Assignments View (eye icon):
General Overview Table
Shows the overall progress of all assignments.
Assignment Details - Click on the eye icon
Displays interaction-level details, including status:Pending
Completed
Failed (with error details)
Progress
Assigned/Expected
In this view, you’ll see a few helpful buttons:
Retry allows you to run the router again, so it can attempt to find additional interactions that match the expected criteria.
You’ll also see the Router Info button, where you can review the router’s configuration, including its type, assigned users or teams, frequency, and other settings.
Additionally, there’s an Activity Log button where you can check the router’s activity history.
Possible Errors
Interaction Not Found — The interaction no longer exists in the system.
Interaction Already Processed — The interaction has already been processed by the same AI product.
AI Disabled — The AI product has been disabled.
AI Usage Limit Reached — The monthly limit for the AI product has been exceeded.
No Agent Comments — The interaction contains no agent comments.
Max Attempts Reached — The router attempted to process the interaction 5 times unsuccessfully.
Unknown Error — An unexpected issue occurred. Please contact support.
Replace Auditor - Routers Reassignment -
In Oversai, you can reassign QA interactions from one grader to another. This is especially useful when a grader is unavailable (e.g., on vacation or sick leave), ensuring evaluations continue without disruption.
You’ll find the reassignment option available at different levels within Routers, depending on how granular you want the change to be:
1. Global Level (Across Multiple Routers)
At the highest level, you can reassign all interactions currently assigned to a specific grader across multiple routers.
Example:
Reassign all interactions from Router A and Router B from one grader to another in a single action.
2. Router Level (Specific Router)
At this level, you can reassign all pending interactions within a specific router, optionally filtered by execution dates.
Example:
Reassign all pending evaluations in Router A.
3. Execution Level (Single Router Run)
At the most granular level, you can reassign interactions from a specific router execution.
Example:
Reassign only the pending interactions from the March 8th execution of Router A.
Credit Consumption
Router Mode: Credits are allocated monthly, per AI product.
Important:
Router credits are not limited daily, since usage depends on router configuration.
Once the monthly credit limit for an AI product is reached, all router assignments for that product will fail with the error:
AI usage limit reached.Contact our support team if you need to increase your monthly credit limit.
📌 Summary
Routers in Oversai automate how interactions are selected, distributed, and processed across your quality assurance and AI workflows.
Instead of manually assigning evaluations or running AI analysis, routers allow teams to define rules that automatically sample interactions, assign them to reviewers, or send them through AI processes on a schedule.
Start simple, monitor results, and adjust over time.
A well-configured router becomes a critical tool for maintaining CX quality at scale.
Still need help?
Contact us