AI Is Getting More Human: What the HBR Data Actually Shows

A 2025 Harvard Business Review article by Marc Zao-Sanders of Filtered.com found that the top use case for generative AI is no longer code generation or task automation. It is therapy and companionship. Organizing life and finding purpose ranked second and third. Personal and professional support now accounts for 31% of all AI use, nearly doubling from 17% in 2024.

Something I am starting to observe in conversations with business owners, employees, and students across the Philippines: people are using AI for things nobody put on the official list.

Not the company rollout plan. Not the department chatbot. Something quieter. Someone processing a hard week. Someone asking for help sorting out their priorities. Someone just talking, at 2 a.m., because the AI is there and it listens.

I came across a 2025 Harvard Business Review article by Marc Zao-Sanders, co-founder of Filtered.com, that put data behind what I had been noticing. His team analyzed thousands of Reddit and Quora posts and found something that stopped me.

The top use cases for generative AI in 2025 are not technical.

They are human.

The Data That Changes the Conversation

Zao-Sanders published “How People Are Really Using Gen AI in 2025” on April 9, 2025. It was an update to his widely shared 2024 research. What shifted between the two years is significant.

In 2024, the top use was generating ideas. Therapy and companionship was second. Technical uses still dominated the top ten.

In 2025, therapy and companionship moved to first place. Two entirely new use cases entered the top three: organizing my life (#2) and finding purpose (#3). Both were not even tracked in 2024.

Personal and professional support, the category that holds all three of these, went from 17% of total AI usage in 2024 to 31% in 2025. That is not a small movement. That is a category nearly doubling in one year.

Zao-Sanders summarized it this way: most experts expected AI to prove itself first in technical areas. While it is doing that, the research suggests AI may help people as much or more with their human needs.

What People Are Actually Using AI For

The three leading use cases in 2025 tell a specific story.

Therapy and companionship (#1) captures people using AI to process grief, work through anxiety, manage loneliness, and get emotional support at any hour, without judgment, often for free. The Philippines has about one mental health clinician for every 800 people, according to estimates cited in local health coverage. AI is not a replacement for professional care. But it is filling a gap that nothing else is filling at scale.

Organizing my life (#2) is about structure. Setting habits. Reflecting on intentions. Taking small steps forward at home and at work. This is not an enterprise feature request. This is a person on a Sunday night trying to get a clearer picture of their week.

Finding purpose (#3) is the one that surprised me most. People are using AI to define their values, move past roadblocks, and make progress on personal development. Problem reframing. Next-step planning. Work that would typically require a life coach or a mentor — available to anyone with an internet connection.

Further down the list, people are using AI to boost confidence (#18), have deep conversations (#29), and even attempt to connect with deceased loved ones (#33).

These are not productivity metrics. These are human needs.

Why This Matters for Filipino Business Leaders

If you are running a team, managing an organization, or thinking about how AI fits into your operations, this data is worth sitting with.

Your employees are already using AI. Not just for the tasks you assigned. They are using it for things that are personal, emotional, and often invisible in any formal adoption report.

At PAIBA and through the programs we run at Olern, I see this regularly. Business owners who attend AI literacy sessions often arrive focused on workflow efficiency. By the end of the session, the questions shift. Someone asks whether AI can help them think through a hard decision. Someone else mentions they have been using it to process stress. Another person says they use it to rehearse difficult conversations before they happen.

The concern some parents raise — that their children might form emotional attachments to AI — is not as far-fetched as it sounds. It is already happening, across age groups, across demographics, in ways that are both understandable and worth watching carefully.

What You Can Do With This Information

This is not an argument to slow down AI adoption. It is an argument to understand it more accurately.

Expand your definition of AI use. Most organizational AI audits track tool logins, task completions, and prompt counts. Those metrics miss everything in the top three. Start asking better questions: Is AI helping your team manage stress? Is it helping people organize their thinking? Do employees feel supported by it, or are they using it as a coping mechanism for something the organization should be addressing directly?

Do not treat emotional AI use as a problem to solve. For many people in your organization, AI is giving them access to reflection, structure, and support that they did not have before. That is generally a good thing. Blanket restrictions on personal AI use at work can remove a resource that employees have quietly come to rely on.

Watch for dependency signals in your own team. There is a difference between someone using AI to organize their thinking and someone using AI because the workplace itself is isolating or overwhelming. If AI is filling gaps that should be filled by management, culture, or adequate mental health support, that is the thing to address — not the AI.

Start with a conversation, not a policy. If you want to understand how people in your organization are actually using AI, ask them. The HBR data confirms that the real use patterns look very different from the official ones. A short, honest conversation reveals more than any usage dashboard.

Where to Start This Week

Pick one person on your team who uses AI regularly. Ask them one question: what is the most personal way you have used AI this month?

You will likely hear something that was not in your adoption plan. That answer is your next data point.