There is an entire industry built around teaching support agents to be more empathetic. Workshops. Scripted phrases. Coaching scorecards with an "empathy" column. The message, implicit in most of it, is that empathy is a trainable competency: a skill on a scale from one to ten that you can move upward with the right curriculum.
I think this framing is wrong, and it quietly costs support teams more than they realize.
Not because empathy doesn't matter in support work. It matters enormously. But the mistake is treating it as something to develop rather than something to screen for. Empathy, genuine care for the person on the other end of the ticket, is not a skill you build in someone who doesn't already have it. It is a baseline requirement for the role, the same way calm under pressure or clear written communication is. You either bring it in on day one or you don't. What you can teach is the language and the habits that help someone who already cares express that care clearly.
The Case for the Other Side
I want to engage with the counterargument honestly, because it isn't weak.
Emotional intelligence varies. Someone can genuinely care about a frustrated customer but lack the vocabulary to say so under pressure, especially in writing, especially across a language or cultural gap. Structure and coaching can help that person find the register. Teaching someone to slow down, to name the customer's frustration explicitly, to resist the instinct to immediately defend the product: that is real and learnable, and some of it falls under what teams call empathy training.
There is also a version of this training that is really about tone and de-escalation. That is useful. Teaching agents how to phrase a policy decline without sounding bureaucratic, how to give bad news without triggering an escalation, how to close a ticket so the customer feels heard even when the answer is no: all of that is worth investing in.
I don't disagree with any of that. My argument is narrower: calling it empathy training obscures what it actually is, and it often distracts from the investments that would move outcomes the most.
What Actually Moves CSAT
I have watched CSAT scores closely across a few different support environments, and the pattern that holds is this: the agents with the consistently highest scores are almost never the ones who sound the warmest in tone. They are the ones who can answer the question correctly, quickly, and without creating extra work for the customer.
Warm tone is an output of capability, not a cause of it. An agent who knows the product well, has system access to actually do something useful, and can resolve most issues in a single interaction sounds warm because they are not stressed and guessing. They can afford to be warm. The agent who has to apologize their way through a problem they can't fix, forward the ticket to someone else, and promise a callback that may not come: that agent can follow every empathy script in the training deck and the customer will still leave frustrated.
The CSAT score reflects whether the customer's problem got solved. Everything else is texture.
So the question becomes: where should the training budget go, if not into cultivating empathy?
Product knowledge, the kind that lets an agent answer directly rather than promising to follow up. System access and authority, the ability to actually resolve what the customer is calling about without three escalation steps. Escalation judgment, knowing clearly when a ticket is yours to resolve and when it isn't, so no time is wasted figuring that out live. Good internal documentation, so agents are working from current information and not from memory or a three-year-old wiki.
These investments make agents more capable. More capable agents produce better outcomes and, as a consequence, naturally warmer interactions.
The Ratio Test
Pull up your support team's training calendar or onboarding curriculum. Add up the hours allocated to empathy, communication tone, and CSAT coaching. Now add up the hours spent on product depth, tool proficiency, escalation procedures, and system access walkthroughs.
If the first bucket is larger than the second, I'd argue the priorities are inverted.
This is not a hypothetical. In a lot of support organizations, especially those that grew fast or shifted to distributed remote teams, the training stack ends up front-loaded with soft-skills content because it's easier to produce and easier to run. A Zoom session on empathetic language is faster to put together than a thorough product walkthrough or a live escalation simulation. So it gets scheduled. And the harder technical onboarding gets abbreviated or left to on-the-job learning, which means the agents who pick it up fastest are the ones who already had relevant background, and the rest are left to improvise.
The ratio test tells you what the organization actually believes about what makes a good support interaction. If the belief, revealed by how time is spent, is that tone is the primary lever, then tone is what gets trained, and capability is what gets assumed.
The best CSAT scores I have seen come from teams that treated capability as the job and let warmth be the byproduct. Confidence is warm. Competence is kind.
The Real Question Is About Authority, Not Attitude
Here is where this lands for me practically. When a support team has a CSAT problem, the first instinct is often to look at how agents are communicating: are they being warm enough, empathetic enough, following the right script? The coaching session gets scheduled, the empathy rubric gets updated, and the scores may nudge up for a quarter before settling back.
The more useful question is: what are agents allowed to do? Can they issue a refund without a manager approval? Can they see the full account history, or just the last five tickets? Can they tell the customer definitively what the fix is, or do they have to hedge because the answer requires checking with someone else?
Authority and access are what determine whether an interaction can actually be resolved. Empathy determines whether the customer feels cared about during an interaction that may or may not resolve. Both matter. But when you're troubleshooting a CSAT problem, the first question to answer is whether agents have the tools and authority to do their job. The second is whether they know how to do it well. Tone comes third.
If your agents have to apologize three times and open a callback case before they can issue a refund, no amount of empathy training is going to move your scores in a lasting way. The script isn't the bottleneck.
I hold this opinion partly from having been on both sides: as someone who has managed support queues and as someone who has called support for things I genuinely needed fixed. The interactions that left me satisfied weren't the ones where someone sounded most sorry. They were the ones where someone knew what they were doing and fixed it.
Empathy is what got them into the job. Capability is what made them good at it.
Push back if you see it differently. I am genuinely curious where this argument breaks down for other people's teams.
If this connects to something you're working through, reach out directly.
Comments