#AI Copilot

#What is AI Copilot?

AI Copilot is your intelligent assistant for call center analytics. Ask questions in natural language and get instant answers about your data. Copilot can also take actions like flagging calls or adding notes.

Think of it as having a data analyst available 24/7 who knows everything about your calls.

Access: Dashboard → AI Copilot (sidebar) or press ⌘K / Ctrl+K anywhere

[SCREENSHOT: AI Copilot interface with chat window and campaign selector]

#Capabilities

#Data Queries

Ask questions and get instant answers:

Question Type	Examples
Metrics	"What's our average score this week?"
Comparisons	"Compare this week to last week"
Filtering	"Show me calls with score below 60"
Aggregations	"How many calls did we process yesterday?"

#Agent Analysis

Get insights about your team:

"Who are our top 5 performers?"
"Which agent has the most failed compliance checks?"
"Show me John Smith's recent calls"
"Who needs coaching?"

#Trend Analysis

Understand patterns over time:

"How has our score changed over the last month?"
"Which checks are failing most often?"
"Is our compliance rate improving?"

#Taking Actions

Copilot can make changes on your behalf:

Action	Example
Flag for review	"Flag this call for review"
Add notes	"Add a note: needs follow-up training"
Bulk flag	"Flag all calls from John with score below 60"
Dismiss reviews	"Dismiss the review on call [ID]"

#How to Use

#Opening Copilot

Three ways:

Click AI Copilot in the sidebar
Press ⌘K (Mac) or Ctrl+K (Windows)
Click the Copilot icon in the top navigation

#Basic Conversation

Type your question in plain English
Press Enter or click Send
Copilot responds with data and insights
Ask follow-up questions to drill down

#Campaign Context

You can scope questions to a specific campaign:

Select a campaign from the dropdown (top of Copilot)
Questions automatically focus on that campaign
Or ask about "all campaigns" explicitly

#Conversation History

Copilot remembers context within your session:

Ask follow-ups like "Show me more details on the first one"
Reference previous answers: "What about for Jane instead?"
Context is maintained as you continue the conversation

Scenario	Context Saved?
Same session	✓ Yes
Page refresh	✗ No, conversation resets
Next day	✗ No, starts fresh
Different device	✗ No, sessions are independent

Tip: Copy important insights before leaving Copilot if you need to preserve them.

#Asking Questions

#Effective Queries

Less Effective	More Effective
"How are we doing?"	"What's our average score this week compared to last?"
"Show agents"	"Show me agents with declining scores"
"Calls"	"Show me calls from yesterday with negative sentiment"

#Query Patterns

Simple metrics:

"What's our average score?"
"How many calls did we process today?"
"What's our pass rate for compliance checks?"

Filtered queries:

"Show me calls with score below 70"
"Find calls from agent John Smith"
"List calls from last week with negative sentiment"

Comparative:

"Compare this week's scores to last week"
"Which campaign is performing best?"
"How does John compare to the team average?"

Trend-based:

"How has our score changed over the past month?"
"Is our compliance rate improving?"
"Show me the weekly trend for customer verification failures"

#Taking Actions

Copilot can perform actions when you ask:

#Flagging Calls

You: "Flag call cmj7abc123 for review because of compliance issues"

Copilot: "I've flagged call cmj7abc123 for review with the reason: 
         compliance issues"

#Adding Notes

You: "Add a note to call cmj7abc123: Agent needs training on verification"

Copilot: "Done! I've added the note to call cmj7abc123"

#Bulk Actions

You: "Flag all calls from today with score below 65"

Copilot: "I found 3 calls matching that criteria:
         - call_001.mp3 (score: 62)
         - call_042.mp3 (score: 58)
         - call_089.mp3 (score: 64)
         
         Should I flag them for review?"

You: "Yes"

Copilot: "Done! I've flagged 3 calls for review with the reason: 
         Low score (below 65)"

#Safety Confirmations

Copilot asks for confirmation before:

Bulk actions (affecting multiple items)
Destructive actions (if any)

This prevents accidental changes.

#What Copilot Can Access

#Data Access

Data Type	Access Level
Call recordings metadata	✓ Full access
Transcriptions	✓ Full access
Check results	✓ Full access
Agent profiles	✓ Full access
Campaign settings	✓ Read-only
Reports	✓ Can summarize
Keyword matches	✓ Full access

#What Copilot Cannot Access

Other organizations' data
Billing information
User passwords or credentials
Audio files directly (works with text only)

#Example Conversations

#Simple Query

You: "What's our average score this month?"

Copilot: "Based on 234 completed calls this month, your average 
         quality score is 82.4. This is up 3.2 points from last 
         month (79.2)."

#Agent Analysis

You: "Which agent has the most calls needing review?"

Copilot: "Sarah Johnson has 8 calls in the review queue, followed 
         by Mike Chen with 5. Most of Sarah's flagged calls are 
         due to verification check failures."

#Multi-Step Investigation

You: "Our scores dropped this week. What's happening?"

Copilot: "I can see average scores dropped from 81.2 last week to 
         74.8 this week (-6.4 points). Let me analyze...
         
         Key findings:
         1. 'Customer Verification' pass rate dropped from 89% to 67%
         2. Most failures are from agents: Sarah J. (12), Mike C. (8)
         3. The drop started on Tuesday
         
         Would you like me to look deeper into any of these areas?"

You: "Tell me more about Sarah's calls"

Copilot: "Analyzing Sarah Johnson's calls from this week...
         
         Compared to last week:
         - Total calls: 24 (up from 18)
         - Average score: 68.2 (down from 84.1)
         - Verification check: 58% pass rate (was 92%)
         - Call duration: 4.2 min avg (was 6.1 min)
         
         Pattern: Her calls are shorter this week and she's 
         frequently skipping the verification step.
         
         Would you like me to:
         1. Flag her low-scoring calls for review?
         2. Show specific examples?
         3. Add a coaching note to her profile?"

You: "Flag her calls below 70 and add a coaching note"

Copilot: "Done! I've:
         ✓ Flagged 9 calls for review (score < 70)
         ✓ Added note to Sarah's profile: 'Week of Jan 15: 
           Verification compliance dropped. Review flagged calls 
           and schedule coaching session.'"

#Available Tools

Copilot uses these internal tools to answer questions:

#get_calls_data

Retrieves call information with filters:

Campaign
Agent
Date range
Status
Score range
Call type

#get_check_results

Gets quality check results:

Check name
Check type
Result (pass/fail/na)
Agent
Date range

#get_call_insights

Retrieves AI-generated insights:

Executive summaries
Sentiment analysis
Key topics

#get_campaigns_summary

Overview of all campaigns with stats.

#get_available_checks

List of quality checks configured per campaign.

#Limitations

Copilot has some boundaries:

Cannot Do	Alternative
Create or modify quality checks	Use the Quality Checks UI
Delete calls or campaigns	Use the respective detail pages
Access other organizations	Not possible
Modify beyond your permissions	Respects your role
Access real-time data	Data may be 1-5 minutes delayed
Play audio	Use the call detail page

#Keyboard Shortcuts

Shortcut	Action
`⌘K` / `Ctrl+K`	Open Copilot from anywhere
`Enter`	Send message
`Shift+Enter`	New line in message
`Escape`	Close Copilot
`↑`	Recall previous message

#Best Practices

#Be Specific

Include time ranges: "this week", "yesterday", "last 30 days"
Name specific agents, campaigns, or checks
Clarify what you're looking for

#Use Follow-Ups

Start broad, then narrow down
Reference previous answers
Let Copilot guide the investigation

#Confirm Actions

Review Copilot's understanding before confirming bulk actions
Check the list of affected items
Add context when flagging

#Combine with Other Tools

Use Copilot for quick questions
Use Reports for formal documentation
Use Call Detail pages for deep review

#FAQ

#Does Copilot use my data to train AI?

No. Your data is used only to answer your questions in real-time. It is not used to train AI models.

#Can Copilot access audio files?

No. Copilot works with text data only (transcripts, metadata, scores). To listen to audio, open the call detail page.

#Why does Copilot say "I don't have access to that"?

Copilot respects your permission level. It also cannot access:

Other organizations
Deleted data
Billing information

#Are conversations saved?

No. Conversations reset when you leave Copilot or refresh the page. Copy important insights before closing.

#Can I ask about specific call IDs?

Yes. Use patterns like:

"Tell me about call cmj7abc123"
"What was the score for call ID [paste ID]?"
"Show check results for [call ID]"

#Is there a limit to questions?

No hard limit, but:

Very complex queries may take longer
Bulk actions have reasonable limits
Extreme usage may be rate-limited

#Can Copilot write quality checks for me?

Copilot can suggest check prompts and help refine wording, but cannot directly create or modify checks. Use the Quality Checks UI for that.

#How current is Copilot's data?

Copilot queries live data with potential delay of 1-5 minutes. Very recent uploads may not appear immediately.

Review Queue - Where flagged calls appear
Agents - Agent performance data
Reports - Automated alternative to Copilot
Quality Checks - Understanding check results

#AI Copilot

#What is AI Copilot?

#Capabilities

#Data Queries

#Agent Analysis

#Trend Analysis

#Taking Actions

#How to Use

#Opening Copilot

#Basic Conversation

#Campaign Context

#Conversation History

#Asking Questions

#Effective Queries

#Query Patterns

#Taking Actions

#Flagging Calls

#Adding Notes

#Bulk Actions

#Safety Confirmations

#What Copilot Can Access

#Data Access

#What Copilot Cannot Access

#Example Conversations

#Simple Query

#Agent Analysis

#Multi-Step Investigation

#Available Tools

#get_calls_data

#get_check_results

#get_call_insights

#get_campaigns_summary

#get_available_checks

#Limitations

#Keyboard Shortcuts

#Best Practices

#Be Specific

#Use Follow-Ups

#Confirm Actions

#Combine with Other Tools

#FAQ

#Does Copilot use my data to train AI?

#Can Copilot access audio files?

#Why does Copilot say "I don't have access to that"?

#Are conversations saved?

#Can I ask about specific call IDs?

#Is there a limit to questions?

#Can Copilot write quality checks for me?

#How current is Copilot's data?

#Related Documentation

Documentation