Key points
- Empathy in customer service is measurable. Peer-reviewed research has scored it from language patterns and vocal pitch since 2014, and the same signals appear in sales and support transcripts.
- Insight7 scored 6,209 sales conversations and found empathy produced the widest single behavioural difference in the dataset: 4.7 for top performers against 2.8 for the bottom tier, a 68% difference.
- The 4 A’s give agents something to do about it: Acknowledge, Align, Assure, Act. Each one shows up in a transcript as a specific phrase pattern a QA rubric can score.
- Insight7’s AI Call Scoring already treats empathy as a configurable QA dimension across 100% of calls, with every score linked to the transcript moment behind it, so you can coach it rather than hope for it.
Your QA scorecard probably measures script adherence, account verification, and handle time. Yet empathy in customer service sits below that, scored out of five by whoever happened to listen, and treated as a trait the agent either has or does not. Well, good news.
Empathy in customer service is measurable, and our own data on 6,209 sales calls shows it separating performance tiers more sharply than any other behaviour we track.
Here’s what the scored conversations showed and example statements you can start using right away.
Can empathy be measured on a sales or customer service call?
Yes. Empathy has been scored objectively from conversation data in peer-reviewed research since at least 2014, using two independent methods.
Language style synchrony
Researchers at the University of Washington and the University of Utah analysed 122 counselling transcripts totalling more than 362,000 words, measuring how closely the two speakers matched each other’s phrasing patterns.
Sessions rated high on empathy showed significantly greater language style synchrony across 11 linguistic categories than low-empathy sessions, at p less than .01, with a substantial effect size of d = 0.62. The finding held even after controlling for reflective listening statements, which means synchrony captured something the obvious verbal cues missed.
Vocal pitch synchrony
A separate study from the same research group extracted vocal arousal measures from 89 sessions and tested whether speaker and listener pitch tracked together. Higher synchrony aligned with higher observer ratings of empathy.
Both methods share one useful property for anyone running QA: they score from the recording, with no observer sitting in and no self-report survey afterwards.
What our own call data shows
We scored 6,209 real sales conversations across 12 dimensions and sorted them into performance tiers. Empathy and rapport produced the widest single behavioural difference in the entire dataset.
| Metric | Top tier | Bottom tier | Difference |
|---|---|---|---|
| Empathy and rapport | 4.7 | 2.8 | +68% |
| Enthusiasm score | +35% | baseline | +35% |
| Exclamation count | 2.4x | baseline | 2.4x |
A 68% difference on a single dimension is larger than what we found on question count, talk ratio, or actionable outcomes. For a CX leader at a 250-person company deciding where coaching hours go, empathy scoring higher than the criteria already on your rubric is a reason to add it to the rubric.
The pattern repeats outside B2B sales. In our high-ticket wellness research, top-tier advisors scored 36% higher on emotional connection than bottom-tier advisors and enrolled 100% of patients where the bottom tier enrolled 33%, working the same programmes at the same prices.
Want to see your own team’s empathy scores across every call? Start free
What predicts empathy scores in real sales and support calls?
Two things predict a high empathy score: whether the agent matched the customer’s language and pacing, and whether acknowledgement arrived at the moment of friction rather than at the end of the call.
Timing carries more weight than volume. In our wellness consultation data, top advisors deployed empathy signals at specific friction points, when the patient raised a cost concern, a health anxiety, or frustration with a previous treatment. They acknowledged before solving.
Bottom-tier advisors delivered the same sympathetic language, arriving after the recommendation, where it read as a formality.
Where empathy sits among your other QA criteria
Four categories cover most call evaluation rubrics:
- Compliance: disclosures, verification, required language
- Communication quality: tone, clarity, empathy, active listening
- Resolution effectiveness: whether the issue was solved and next steps confirmed
- CX signals: sentiment shift across the call, escalation risk, churn language
Weighting differs by team. A support floor handling billing complaints might put 40% of the score on empathy and resolution combined. A sales floor might split that same 40% between empathy and objection handling, since a rep who acknowledges the concern and then fails to address it has done half the job.
But there is a real trade-off to know before you weight empathy heavily on a sales team.
Research from Cambridge Judge Business School separates cognitive empathy, meaning the ability to understand the buyer’s perspective, from emotional empathy, meaning feeling what the buyer feels.
Cognitive empathy improved pricing, sales volume and service quality together. Emotional empathy raised volume while pushing prices down, because reps who felt the buyer’s discomfort conceded on price to relieve it.
What that means for your rubric: score whether the rep demonstrated understanding of the customer’s situation, rather than scoring how warm they sounded. The first predicts revenue while the second can cost you margin.
The 4 A’s of customer empathy
The 4 A’s are Acknowledge, Align, Assure, and Act. Each one produces a phrase pattern you can find in a transcript, which is what makes the framework usable as a scoring rubric rather than a training poster.
- Acknowledge: Name what the customer said before doing anything else. A support agent handling a third callback about the same billing error acknowledges the repetition, not just the error.
Scoring criterion: did the agent restate the customer’s specific issue in their own words within the first 90 seconds?
- Align: Show that the concern makes sense from where the customer sits. Alignment differs from agreement, and it works without conceding fault.
Scoring criterion: did the agent connect the customer’s reaction to their situation, rather than moving straight to policy?
- Assure: Tell them what happens next and who owns it. Uncertainty drives escalation more reliably than bad news does. Scoring criterion: did the agent state a specific next action with a timeframe?
- Act: Do the thing. Empathy without resolution reads as a script, and customers recognise it quickly.
Scoring criterion: was the stated action completed or scheduled before the call ended?
Some versions of this framework use Apologize in place of Align. We prefer Align for sales conversations, where there is often nothing to apologise for and an unnecessary apology weakens the rep’s position. Support teams handling service failures may want both.
What empathy sounds like on a call
Here are five patterns, one per moment, tied to the framework above. Adapt the wording to your brand voice rather than reading them verbatim, since scripted empathy scores worse than none.
- Acknowledging a repeat contact: “This is the third time you’ve had to call about this, and that’s on us to fix properly today.”
- Aligning with frustration: “If I’d been waiting nine days for a reversal with no update, I’d be calling too.“
- Assuring during a long resolution: “I can’t close this today, so here’s what I’m doing instead: I’ll update you every Friday until it’s resolved, even if there’s no news.”
- Acknowledging cost concern in a sales call: “You mentioned the budget review in March. Let me show you what this looks like against that timeline before we talk about numbers.”
- Acting on a health or financial anxiety: “You said you’ve tried two programmes that didn’t hold. Let’s start with why those stopped working before I recommend anything.”
Notice that none of these say “I understand how you feel.” Generic sympathy statements score poorly on language style synchrony because they match nothing the customer said.
A low-empathy call rarely ends in an immediate complaint. The customer simply resolves their issue and privately decides you are difficult to deal with, which shows up later as an escalation, a poor CSAT response, or a renewal that does not happen.
The measurable version of that cost appeared in our wellness data. Bottom-tier consultations degenerated into advisor monologues running 80% or more of the talk time, and those advisors enrolled 33% of patients against 100% for the top tier. Same programmes, same pricing, same patient pipeline.
Our insurance research found the same structure. Top-tier agents generated 16.1 customer turns per conversation against 3.2 for the bottom tier, and closed 65.4% of conversations where bottom-tier agents closed none. A customer who stops talking has stopped participating in the outcome.
Lack of empathy in customer service is expensive because it is invisible. Nobody escalates a call that was merely cold, so it never reaches QA, and the pattern repeats until the churn report arrives with no explanation attached.
See where your calls are missing the moment — Book a demo
How AI scores empathy across every sales and support call
Empathy scoring works the same way any other QA criterion does, applied to every conversation instead of the two or three per agent someone had time to review.
Here is how it runs in Insight7:
- Empathy becomes a configurable dimension on your scorecard, alongside compliance, resolution and communication quality. You define what each score level looks like, so a 4 out of 5 means the agent acknowledged the issue at the friction point rather than meaning the reviewer liked the call.
- Every recording gets scored after syncing from Zoom, Teams, Google Meet, Dialpad, Aircall, RingCentral, Talkdesk or Five9.
- Each score links to its evidence. You can open the exact transcript passage behind an empathy rating, which is what lets an agent inspect a low score rather than absorb it.
- Sentiment shifts get tracked across the call, so you can see the moment a customer’s tone changed and check what the agent said next.
- Low scores generate coaching. Weak empathy ratings feed AI roleplay scenarios built on that specific gap, letting an agent practise a complaint-handling conversation before the next live one.
Fresh Prints ran into the manual ceiling on exactly this. Evaluating high volumes of interviews and customer conversations made timely, personalised feedback difficult to deliver.
Automated scoring of agent tone, sentiment shifts and key conversation moments let agents practise flagged behaviours immediately through interactive scenarios rather than waiting for a scheduled manager session, which removed the supervisor bottleneck without loosening the standard.
Pro tip: Test the idea before configuring a rubric. Our free Sentiment Analysis Tool scans an uploaded conversation for emotional tone and friction points.
Start scoring empathy on every call with Insight7
If you came here for statement examples, take the five patterns above and the 4 A’s framework, and skip the rest. That is a complete answer to the question of how to demonstrate empathy on a call.
But if you lead a sales or CX team and want empathy to become a coached behaviour rather than a hoped-for trait, the difference is coverage.
Odun Odubanjo, Insight7’s founder, framed the goal this way when talking about AI in customer service workflows:
“Your goal should be around freeing up your team to do the work that really matters. Can your team show more empathy to customers? Can they use more judgment in their day-to-day?”
Insight7’s AI Call Scoring scores empathy as one dimension among the criteria you define, across every call, with the transcript evidence attached.
Upload five calls and see how the scores compare to what you expected.
FAQs about measuring empathy in customer service
Can empathy be measured on a sales or support call?
Yes. Peer-reviewed research has scored empathy objectively from conversation data since 2014, using language style synchrony across 11 linguistic categories and vocal pitch synchrony between speakers. Insight7’s analysis of 6,209 sales conversations scored empathy and rapport at 4.7 for top performers against 2.8 for the bottom tier.
What are the 4 A’s of customer empathy?
Acknowledge, Align, Assure, and Act. Acknowledge the specific issue in the customer’s own terms, align by showing the reaction makes sense from their position, assure them of the next step and its timeframe, then act on it. Some versions substitute Apologize for Align, which suits service failures better than sales conversations.
How do you show empathy in customer service?
Name the customer’s specific issue before offering a solution, acknowledge friction at the moment it appears rather than at the end of the call, and confirm a next action with a timeframe. Generic sympathy statements score poorly because they match nothing the customer said, which is measurable through language style synchrony.
