Agents

Agents is where every person who handles your conversations comes into focus. Each agent's calls are transcribed, scored, and aggregated, so you can see who they are, how they're performing over any window of time, and what's actually happening across their interactions.

What you can do

• Browse your complete list of agents with their name, username, email, recording source, and total calls.
• Search agents by name or email and narrow the list with Filter Agents.
• Open any agent to see a full performance profile for a date range you choose.
• Track headline metrics like total calls, average handle time, sentiment, grammar, rate of speech, and pronunciation, each with a trend versus the previous period.
• Read an activity heatmap that colors each day by the measure you care about.
• See Top Events detected across an agent's conversations.
• Review rubric and scorecard scores tied to the playbooks that processed the agent's calls.
• Jump from any profile straight into the agent's recently processed calls and open the full conversation.

How it works

Agents aren't something you add by hand. They appear automatically as your integrations sync conversations. Behind the scenes, RedFlare keeps a dedicated record of each agent and ties it to the calls that person handled, matching on their email so every interaction lands under the right profile.

Two layers of data come together on a profile:

1. Call-level facts come from every interaction, whether or not it has finished processing. This is where total calls, calls today, and average handle time come from, so volume and timing show up even before a call is fully analyzed.
2. Processed metrics come from calls that have run through a playbook. RedFlare separates the agent's speech from the customer's, then scores the agent's side for sentiment, rate of speech, grammar, and pronunciation. Those per-call scores are averaged across your selected window to produce the numbers on the profile.

For every metric, RedFlare also computes the same figure for the immediately preceding period of equal length, which is what powers the up or down trend you see on each card. The same scores are bucketed by day to build the heatmap, rolled up by event type to build Top Events, and grouped by playbook, rubric, and scorecard to build the scores section. The organization-wide agents dashboard is refreshed on a rolling schedule (roughly every 15 minutes), so leaderboards and trends stay current without you doing anything.

note

A few headline numbers, like impact score and the org-level sentiment trend, are summary KPIs computed across all of an agent's processed calls rather than tied to a single conversation.

Browse and find agents

The Agents page opens to a searchable, sortable list of everyone whose conversations are being analyzed.

1. Open Agents from the sidebar.
2. Use the Search by name or email box at the top to find a specific person.
3. Select Filter Agents to narrow the list by fields such as name, email, recording source, username, or total calls.
4. Click a column header to sort, and use the pagination controls to move through long lists.
5. Hit the refresh button anytime to pull in the latest synced data.

If you've searched or filtered and nothing matches, you'll see No Agents found with a quick way to clear your search or filters and get back to the full list.

note

If the page is completely empty, no agents have been synced yet. You'll see "Create an integration and sync conversations to start seeing agents here" along with a Go to Integrations button to set up your first connection.

Open an agent's profile

Selecting an agent opens their detail screen, a single view that brings together everything known about their conversations.

1. From the Agents list, click the agent you want to inspect.
2. At the top you'll see their profile header with avatar, username, email, and recording source.
3. Use the date range picker in the header to set the window you want to analyze. Every metric, chart, and score on the page updates to match.

If an agent has no calls in the selected window, you'll see No metrics found with a note that there are no calls over the selected filters. Try widening the date range.

Read performance metrics and scores

The profile is built to answer "how is this agent doing?" at a glance, then let you dig deeper.

1. Start with the metric cards row at the top. Each card shows a key number, such as total calls, average handle time, sentiment, grammar, rate of speech, or pronunciation, along with its trend versus the previous period.
2. Scan the activity heatmap to see daily patterns. Switch between Sentiment, Grammar, Rate of Speech, and Pronunciation to color the calendar by the measure you care about.
3. Check Top Events to see what's being detected most often across this agent's conversations.
4. Scroll to the scores section to review rubric-based evaluations. Pick a playbook and rubric to see Scores Overview and Score Analysis for the period.

tip

Watch the trend chip on each metric card. It compares the current window with the period of equal length right before it, so a green change means the agent is improving and a red change flags where attention may be needed.

note

The scores section only shows playbooks linked to the agent's recording source. If you see No Playbooks Available or No Rubrics Configured, no scoring is set up for that source or no rubrics ran in the chosen date range.

Review recent calls

Every profile links straight back to the underlying conversations, so you're never more than a click from the source.

1. On an agent's profile, select View recent calls in the top-right of the header.
2. The Recently Processed Calls panel opens, listing the agent's latest calls with timing, a short summary, and a sentiment indicator.
3. Click a call to open the full conversation and review what actually happened, including the transcript and per-speaker analysis.

If nothing shows up here, you'll see No calls found with a note that no calls are processed yet for this agent.

Tips

tip

Use the date range picker as your main control. Set it once on an agent's profile and the metric cards, heatmap, top events, and scores all stay in sync with that window.

note

Agents appear here automatically as your integrations sync conversations. To bring in more people or sources, head to Integrations rather than adding agents by hand.

tip

When comparing agents, sort the main list by Total Calls first. It's the quickest way to separate high-volume agents from those with too little data to draw conclusions from.

note

Volume metrics like total calls and average handle time appear as soon as a call syncs, but sentiment, grammar, rate of speech, and pronunciation only fill in once the call has finished processing through a playbook. A new agent may show calls before their quality scores catch up.