Conversation Metrics & KPIs
When you open the "Dashboard" page, you’ll see several boxes at the top with numbers. These are the key metrics:
- Total Conversations: This number tells you how many calls or chats your AI agents have handled in total. For example, if it says "150," it means your AI has talked to customers 150 times. If this number suddenly goes down (e.g., from 150 to 50), it might mean there’s a problem with your AI, and you should check the "Details" section to see what’s going on.
- Unique Users: This shows how many different customers your AI has talked to. It counts each customer only once, even if they called multiple times. For example, if the same customer calls three times, they’re still counted as one unique user. If you see 100 total conversations but only 20 unique users, it means many customers are calling more than once—maybe they’re not getting the help they need the first time.
- Total Conversation Time: This is the total amount of time your AI has spent talking to customers, in minutes or hours. For example, if it says "300 minutes," that’s 5 hours of talking. If this number is very high compared to the number of conversations, it might mean your AI is taking too long to help customers.
- Average Handling Time (AHT): This tells you the average time your AI spends on each call or chat. For example, if the AHT is "2 minutes," it means each conversation takes about 2 minutes on average. If the AHT is very high (e.g., 10 minutes), it might mean your AI is struggling to answer questions quickly. You can go to the "Details" section to read the conversation transcripts and see why it’s taking so long.
Below the metrics, you’ll see some charts or graphs. These are called visualizations, and they help you see patterns in your AI’s performance:
- Timelines: This is a line graph that shows how many conversations happen over time. You can choose to see it by day (e.g., May 1, May 2, May 3) or by week (e.g., Week 1, Week 2). For example, if you see a big spike on May 3 with 50 conversations, but only 10 on May 4, it means May 3 was a busy day. You might want to have more AI agents ready on days like that.
- Distributions by Day and Hour: This chart shows which days and times are the busiest. For example, it might show that most calls happen on Mondays at 10 AM. If you know this, you can make sure your AI agents are ready to handle more calls at those times.
Tip: Use these charts to plan your AI’s schedule. If you know Mondays are busy, you can make sure your AI is fully prepared to handle more calls on that day.