2025 · WE TELL IT HOW WE FEEL
DELTA · LARGER THAN IT SOUNDS
From Tasks
to Therapy.
The top use cases for generative AI stopped being technical.
Almost nobody predicted which ones replaced them.
In 2023 We Told It What to Do.
Now We Tell It How We Feel.
That shift is more profound than it sounds, and it did not surprise me as much as it might have. Some reading I did back in 2021 on the trajectory of technological change left me fairly certain this would eventually find its way inside the house.
What I did not predict was the order. When Harvard Business Review published its look at how people were really using generative AI in 2025, the ranking had rearranged itself completely. The top use cases were no longer technical. They were personal.
The List
Nobody Forecast.
The machine still writes the emails. But more people are now opening it to process grief, reframe a goal, build a routine they can actually hold, or make some meaning out of mental clutter that has been accumulating for years.
Always On.
Never Judges.
In 2023 the value was speed and automation. In 2025 it is support, reflection, and something closer to emotional utility. Millions are leaning on these systems for what a therapist, a coach, or a close friend might once have provided.
The reasons are not mysterious. It is available at 2 a.m. It does not judge. And in a great many cases it is good enough to help someone move forward by morning.
I want to resist the urge to moralize about that, because the honest comparison is not between AI and excellent care. It is between AI and expensive, possibly biased care from an in-network provider with a six-week wait that a person cannot afford anyway.
If You Build,
This Is Your North Star.
The advantage is moving from performance to presence. The winners will not simply be the fastest models. They will be the most emotionally attuned, the most context-aware, and the most human in tone.
Emotive design is the new frontier. Cold interfaces are finished. Products that behave like a thoughtful collaborator will take both the consumer and the enterprise stack.
Making people feel seen will outcompete making them move faster. AI that clarifies purpose beats AI that saves ten clicks, because people want more than answers. They want attentiveness, and they want to be trusted with the truth.
We Thought It Would Help Us Do More.
It Is Helping Us Be More.
I know this personally. The hours this technology has given back have not gone into producing more. They have gone into looking inward, which is the only reliable way I have found to look outward well.
That may be the most consequential outcome available here, and it is nearly the opposite of what was advertised. The most powerful thing this technology does might not be to outthink us.
It might be to help us reconnect with what matters.
Let’s build for that.