The Agent Metrics tab shows the value each AI agent generates: how many conversations it handles, how many it moves forward without help, how much time it saves your team, and how its conversations end in each state of your Sales Brain (Playbook).
1. Where to find it
- Go to Artificial Intelligence Agents and open the agent you want to analyze.
- At the top you'll see three tabs: Conversation History, Agent Metrics and Test Sessions.
- Click Agent Metrics.
2. Choosing the period
At the very top you have the Period selector. You can pick a quick period (Today, Yesterday, This week, Last week, This month, Last month, Last 2 months, Last 7, 14 or 30 days) or a Custom range of up to 6 months.
- The figures on the tab are recalculated for the period you choose.
- Each metric is compared with the previous period of the same length (for example, "+24% vs prev. period").
- If you choose a custom range, that comparison isn't shown, because there's no equivalent previous period.
3. AI impact (illustrative reference)
The first block, AI impact this period, gives a visual summary of the kind of savings AI agents bring: cost saved, hours saved, full-time equivalent (FTE) and contacts attended out of hours.
Expanding With AI vs. without AI shows a table comparing the same volume of contacts handled by a human team versus a team with AI agents (first response time, coverage hours, conversations resolved without a human, hours needed and staff cost). The estimate uses a reference rate of âŹ25/h.
Important: the figures in this block are an illustrative example, the same for every account. They are not your agent's real data and they don't change with the selected period. Your real data is in the metrics and charts in the following sections.
4. Main metrics
These four metrics do reflect your agent's real activity in the selected period:
- Conversations managed by AI: the agent's conversations that closed in the period.
- Conversations autonomously advanced: conversations the agent closed on its own, meeting its objective, without anyone on the team stepping in.
- Human time saved: an estimate of the time your team would have spent writing the messages the agent sent (2 minutes per message). It doesn't count conversations where a team member took over.
- AI-assisted conversions: conversations where the agent's objective was met.
Test conversations (the ones in Test Sessions) aren't counted in any metric.
5. Sales Brain (Playbook) coverage
This notice tells you in how many states of your Sales Brain (Playbook) the agent works, for example: "AI covers 3 of 5 states". A state is covered when the agent has an active objective in it. States without coverage are outside the agent's current scope.
The Expand coverage button takes you to the agent's settings, where you can add objectives in more states.
If the agent has no Sales Brain (Playbook) linked, this notice doesn't appear.
6. Outcomes and progression by state
The last section has two views, which you can switch between with the Chart and Table buttons.
Chart: AI conversation outcomes
One bar per state of the Sales Brain (Playbook), plus a Total row, showing how the agent's conversations end:
- Success: the agent closed the conversation meeting its objective.
- Objective not met: the agent closed the conversation without meeting the objective.
- Discarded: the agent closed the conversation and discarded the contact.
- Expired: the conversation closed due to inactivity, because the contact didn't reply, or because the reply limit was reached.
- Out of credits: the conversation closed because there were no AI credits left.
- Moved by a human: a team member took over the conversation.
- Other: other closures, for example when the contact changed state during the conversation.
Hover over each segment to see the exact number of conversations and its percentage.
Table: AI progression by state
For each state of the Sales Brain (Playbook) you'll see:
- Inflow: conversations that started in that state.
- AI advances: conversations the agent moved forward on its own, meeting its objective.
- Human-assisted: conversations where a team member stepped in.
- Autonomous rate: the percentage of inflow the agent moved forward on its own. It's shown in green from 50% up and in amber below. If a state has no inflow, a dash (-) appears.
You can search for a state by name and sort the table by any column.
If the agent doesn't have any active objectives yet, the section opens directly in the Table view.
Conclusion
The Agent Metrics tab lets you see, with real data for the period you choose, how much work each AI agent takes on, in which states of your Sales Brain (Playbook) it performs best, and where it needs help from your team. Use it to decide where to expand the agent's coverage and where to adjust its objectives.
Relevant scenarios
- High volume of incoming contacts: check Conversations managed by AI and Human time saved to see how much load the agent takes on compared with your team.
- A state with a low autonomous rate (amber): check in the chart whether conversations end as "Expired", "Objective not met" or "Moved by a human", and adjust the agent's objective in that state.
- Many "Out of credits" conversations: review your account's AI credit usage so the agent doesn't leave conversations unattended.
Was this article helpful?