xMetrics are evaluagent's AI-predicted customer experience scores. Instead of waiting on survey responses from a small slice of customers, you get a score for every conversation, with reasoning and evidence to back it up.

This guide covers what each xMetric measures, when to use it, and where it appears in the platform.

To see xMetrics in the platform, you need:

The xMetrics features must be enabled for your organisation. Some xMetrics (xNPS, xRepeats, xResolution, xVulnerability) have their own visibility and processing feature flags; others (such as xCSAT and xCES) are configured as part of your AI metrics setup. Contact your evaluagent administrator if you can't see metrics you'd expect.

xCSAT analyses conversation content and predicts how satisfied the customer was, on a scale of 1 to 5. It picks up on gratitude, frustration, tone, and whether the issue was sorted.

When it's useful: When you want satisfaction data on every conversation, not just the ones where the customer fills in a survey. Surveys skew towards extremes; xCSAT covers everyone.

Where you see it: In the Insights sidebar on a conversation, on dashboard widgets (trend and distribution), and as a filter on the Imported contacts list.

xNPS predicts whether the customer would be a Promoter, Passive, or Detractor based on the conversation. You get a numeric score and a category, plus reasoning and evidence.

When it's useful: When you want NPS-style data without survey fatigue. Good for tracking loyalty signals at scale.

Where you see it: Insights sidebar, xNPS dashboard widgets, and as a filter for finding all Detractor (or Promoter) conversations.

xCES predicts how hard the customer had to work to get their issue dealt with. Result is Easy, Neutral, or Difficult, plus an xCES Driver that calls out what most influenced the score (transfers, repeated questions, complex process navigation, and so on).

When it's useful: High-effort experiences are a leading indicator of churn. xCES helps you find friction in your process before customers leave.

Where you see it: Insights sidebar (with the driver), dashboard widgets, and as a filter on the conversations list.

xResolution predicts whether the customer's issue was actually addressed during the conversation. Resolved, Not resolved, or N/A if there isn't enough content.

When it's useful: Combine with xRepeats to measure true first-contact resolution. Resolved + Not a repeat = a clean fix.

Where you see it: Insights sidebar, dashboard widgets, conversation filters. xResolution is processed as part of xNPS analysis, so the two come as a pair.

xRepeats spots when a customer mentions they've contacted you before about the same issue. It looks for explicit phrases like "I called last week about this" or "this is the third time I'm contacting you".

It does not count transfers within the same call or general mentions of being a customer.

When it's useful: High repeat rates point to issues that aren't being properly resolved on first contact. Useful for FCR tracking and root cause work.

Where you see it: Insights sidebar, dashboard widgets, conversation filters.

xVulnerabilities flags conversations where the customer shows signs of needing extra support. It detects categories including:

When it's useful: Many regulated industries require special handling of vulnerable customers. xVulnerabilities gives you an audit trail and helps target coaching where it matters.

Where you see it: Insights sidebar, dashboard widgets, conversation filters.

Open a conversation. The Insights sidebar shows xMetric results under the xMetrics section. For each one you'll see:

If multiple pieces of evidence exist, click View evidence repeatedly to cycle through each one.

xMetrics are predictions based on conversation content. They can't see what happened after the call ended, and they don't know external context. If you see a result that doesn't match what actually happened: