Welcome to this week's issue of Unpacking Meaning. If you received this from a friend and enjoy it, subscribe here.

When AI learns to empathize… what’s left for us?

The last couple of years, marketers have reassured themselves with the same line: "AI can crunch data, but it can't empathize. It can't truly understand what makes people tick."

I'm not buying it anymore.

I recently interviewed Sam Woods for my podcast (episode dropping soon), and he told me about AI agents that literally call competitors, walk through their sales journeys, and report back what they experience.

If AI can do that today, how long before it's simulating empathy? Before it's modeling motivations and mapping why people act the way they do? Spoiler alert: this is already happening too. Maybe not with a 100% accuracy yet, but synthetic research is “growing up”.

I call this the “noticing gap”. It’s the space between what machines can quantify and what humans can perceive. And it’s shrinking fast.

AI is already detecting emotional undertones in transcripts, surfacing the "why" behind patterns of behavior, and building psychological models that look eerily close to empathy.

Which means the edge isn't in noticing anymore. It isn't even in empathizing. Those are becoming commoditized.

The edge will be in knowing what information to collect (not all signals are worth saving), how to structure it into positioning and messaging strategy, and using context well when writing copy, whether with AI or on your own. Context is what turns raw empathy into meaning.

AI can notice. But it still can't decide. It can't choose which signal matters, how to weigh tradeoffs, or which story your company will commit to.

That's our job.

And as more teams spin up agents to do stuff for them and generate insights, the sharper our judgment, framing, and strategic skills will need to get.

Of course, there are limits worth naming.

Studies show AI-written responses to customer complaints or health concerns are often judged as more "compassionate" or "supportive" than human responses.
But people trust it less once they know it's AI. Identical messages are rated as less empathetic when labeled "machine." We still value the credibility of human care.
And empathy without experience is only projection. What looks like understanding may just be pattern recognition. It's the "ELIZA effect," we project meaning onto words that never carried feeling to begin with.

Which raises an important point: AI might close the noticing gap, but it doesn't remove the need for human discernment. In fact, it increases it.

Because if machines can notice everything, the real skill becomes knowing when to let AI handle the response, when to bring in a human, and how to build systems that feel personal even at scale.

AI is beating us at noticing. But the advantage isn't gone. It just moved.

The edge is in strategy, context, and judgment. And that's where we'll find out who actually understands customers, because spotting signals is easy. Deciding what they mean, and what to do with them, is hard.

DISCOVERY

On the topics of context and judgement...

On my recent appearance on the Insights Unlocked Podcast (by UserTesting), I discussed two principles that shape high-impact messaging.

Assumptions undermine clarity. Without direct customer research, teams risk building their messaging on shaky ground.
How you organize insights matters as much as the research itself. My “iceberg framework”, separating what customers say, do, and believe, helps teams distill feedback into actionable, resonant copy. For more, you can catch the full episode of the Podcast.