Skip to main content
The General view shows how your operation is performing overall. The By Agent view shows which agents are driving those results and where differences in performance are coming from. Open the By Agent tab to see a table listing every agent in your workspace. Each row includes key metrics such as:
  • agent name,
  • total calls,
  • successful calls,
  • success rate,
  • average duration,
  • average latency,
  • and average cost.
This is where you move from aggregate reporting to agent-by-agent comparison.

​
The agent table

By default, the table is sorted by call volume, so the busiest agents appear first. Each column helps you evaluate a different part of performance. Success Rate shows how often calls result in live conversations or successful outcomes, depending on the workflow. Large differences between agents are usually caused by one of three things:
  • the quality of the contact list,
  • the calling windows being used,
  • or caller ID and number reputation.
A low success rate does not always mean the agent itself is weak. In many cases, it reflects the conditions the agent is operating in. Average Duration shows how long connected conversations tend to last. Shorter durations can be normal for fast qualification or routing use cases. Longer durations may reflect more involved conversations, but they can also indicate that the agent is rambling, looping, or struggling to close. The number only becomes useful when viewed in context. Average Latency is a technical performance indicator. It measures how long the agent takes to begin responding after the contact finishes speaking. Lower latency generally feels more natural. If one agent’s latency is noticeably higher than others, the cause is often the voice stack configuration. Average Cost helps you compare the efficiency of different agents, especially when they use different voice or model combinations. This is useful when balancing performance against operating cost.

​
Expanding an agent

Click the dropdown arrow on any row to expand that agent’s detailed performance view. This opens the same style of reporting available in the General dashboard, but filtered to that one agent only. You can review:
  • total calls,
  • success rate,
  • average duration,
  • average latency,
  • average cost per call,
  • daily call volume,
  • sentiment,
  • and call outcomes.
This is where deeper diagnosis happens.

​
Comparing agents

The real value of the By Agent view is comparison. When two agents are running similar campaigns but performing differently, this view helps you isolate what is changing the result. In many cases, the answer is in the configuration. For example:
  • if one agent has lower latency and better sentiment, the voice stack may be the reason,
  • if two agents use the same stack but one has longer conversations and more successful handoffs, the prompt or call logic may be stronger,
  • if one agent has a much lower success rate, the issue may be list quality, timing, or caller ID rather than the agent itself.
Comparing agents side by side makes these differences easier to spot.

​
Best practice

A useful way to test changes is to clone an existing agent, adjust one variable, and run both versions under similar conditions for a few days. Then compare them in the By Agent view. This gives you a clean way to test changes to prompts, voice stacks, settings, or logic without guessing which version performs better.