Best practices to calculate CoPilot Conversation Assist overall usage rate?

Julia Z

Hi everyone!

I’m responsible for determining, calculating, and tracking KPIs for agent adoption of LivePerson CoPilot’s Conversation Assist tool.

We want to monitor adoption and ROI. We have the eventual goal of adding Conversation Assist KPIs to agents' monthly scorecards.

Currently, I’ve been measuring the overall usage rate using this formula:

Method 1: Conversations-Based Usage Rate Number of conversations with recommendations used ÷ Number of conversations with recommendations offered × 100

Example (week ending 3/22):

• Conversations with recommendations offered: 8,484
• Conversations where recommendations were used: 21
• Usage rate: 0.2% (21 ÷ 8,484 × 100)

Alternatively, I could calculate it this way:

Method 2: Recommendations-Based Usage Rate Number of recommendations used ÷ Number of recommendations offered × 100

Example (same week):

• Recommendations offered: 1,457
• Recommendations used: 25
• Usage rate: 1.7% (25 ÷ 1,457 × 100)

Each approach gives a slightly different result. I'm curious—how do others here define and track Conversation Assist usage in your organizations? Is there an industry standard or best practice you follow? Any insights would be greatly appreciated!