According to an April 2025 Harvard Business Review article by Marc Zao-Sanders, the top use case for generative AI is therapy and companionship, not productivity or coding. The majority of the top-ten uses are personal. For business leaders, this is a signal: AI adoption may not begin with company policy. It may already be underway in people’s everyday lives.
Something I’m starting to observe in how people talk about AI adoption.
The assumption is that it spreads through organizations. A company launches a policy. Teams get trained. Tools get rolled out. And slowly, people adopt.
That’s one model. But there’s data suggesting a different one is already underway.
In April 2025, Marc Zao-Sanders, co-founder of Filtered, published a piece in Harvard Business Review analyzing how people actually use generative AI. The findings came from Filtered’s 2025 Top-100 Gen AI Use Cases Report, which tracked real usage patterns across platforms including ChatGPT, Copilot, and Gemini.
The chart that circulated from that article tells a different story from what most business conversations assume.
Therapy and Companionship Is the Top Use Case
Not reports. Not code generation. Not automation.
The single most common use of generative AI in 2025 was therapy and companionship. Zao-Sanders grouped them together, he explained, “because both fulfill a fundamental human need for emotional connection and support.”
That’s worth sitting with.
People are not going to ChatGPT to optimize a workflow. They are going to it to feel heard. To process something difficult. To talk through a decision without judgment. In a country like the Philippines, where the Philippine Mental Health Association noted in 2023 that there is less than one mental health worker for every 100,000 Filipinos, that says a lot about the gap people are filling on their own.
Finding purpose and organizing one’s life also appear in the top three. Generating ideas and enhanced learning are there too. Fun, creativity, general advice, and healthier living round out the list.
At least six of the top ten use cases in the report are personal or non-work-related. Coding for professionals does appear on the list, but it sits well below the emotional and personal categories.
The Adoption Story Nobody Is Tracking
For years, the conversation around AI adoption has centered on organizations. Leadership buy-in. Change management. Governance frameworks. Training programs.
All of those matter. We work on exactly those questions through PAIBA and through the workshops we run at Olern. They are not wrong.
But the HBR data points at something that runs underneath all of it. People are already building a personal relationship with AI. They are using it to think, feel, ask, reflect, play, and sometimes just pass time. That behavior is not waiting for an organizational policy to permit it.
And that may explain something about why adoption is accelerating faster than most forecasts predicted. Personal use lowers the barrier. A person who has spent three months talking to an AI about a stressful decision at home arrives at a workshop already past the skepticism stage. They have already crossed the line from “what is this thing” to “I know how this thing works.”
In our sessions with business leaders across the Philippines, we see this regularly. The people who are most comfortable with AI tools are often not the ones who received the most formal training. They are the ones who started using AI for something personal, something low-stakes, and built comfort from there.
Why This Matters for Business Leaders
The implication is not that companies should stop building AI strategies. The implication is that the soil they’re planting into may already be more fertile than they realize.
If employees are already using AI personally, then the question is not how to introduce them to the technology. The question is how to channel an existing habit into a work context.
That is a different problem. And it calls for a different approach.
A training program that teaches a skeptic to use AI is one kind of intervention. A program that helps someone who already uses AI personally to apply that same instinct at work is a much easier conversation. The goal is not to start from zero. The goal is to bridge what people already do on their own into what the organization needs.
Filipino workers, especially younger professionals, are already among the most active AI users in Southeast Asia. They are using AI to draft messages, look up information, think through problems, and yes, to process emotions and find advice. That behavior does not disappear when they clock in. It goes underground, into personal devices, unless the organization creates a visible space for it.
What Business Leaders Can Do With This
Start by asking your team what they already use AI for personally. Not in a policy context. Just curiosity. The answers will tell you more about your team’s actual AI readiness than any formal assessment.
Build a bridge from personal to professional. If someone on your team uses AI to draft messages at home, show them how the same habit applies to work correspondence. The skill transfers. The habit is already there.
Treat comfort, not capability, as the real adoption bottleneck. Most people do not fail to adopt AI because the tools are too complicated. They fail because the work context does not feel as safe as the personal one. Reduce the social risk of trying. Make it visible that leadership uses these tools imperfectly and learns along the way.
Watch for shadow AI before it becomes a liability. If employees are already using personal AI tools for work tasks without disclosure, that’s not necessarily defiance. It’s often an unsupported habit looking for a home. Build policy around what’s already happening.
If your team has not yet been introduced to structured AI thinking for business, PAIBA’s training programs and the Olern platform offer practical frameworks built for Filipino business leaders and professionals.
Frequently Asked Questions
Where did the data on top Gen AI use cases come from?
The data comes from Filtered’s 2025 Top-100 Gen AI Use Cases Report, written up by Filtered co-founder Marc Zao-Sanders in an April 2025 Harvard Business Review article. It tracked actual usage patterns across major AI platforms including ChatGPT, Copilot, and Gemini to identify how people apply generative AI in practice.
Why is therapy and companionship the top use case for AI?
Marc Zao-Sanders grouped therapy and companionship together because both serve the same underlying need: emotional connection and support. People turn to AI for low-judgment conversations, reflection, and processing difficult experiences. In markets like the Philippines, where mental health professional availability is very limited, the appeal is particularly strong.
Does this mean AI is not being used for work?
It is being used for work, but the data shows that personal use cases dominate the top of the list. Work-related uses like code generation and idea generation do appear, but they sit below emotional and personal categories. The broader takeaway is that AI adoption is not exclusively a professional behavior, and personal use may be building the comfort that drives professional adoption.
What does this mean for AI adoption in Philippine businesses?
It suggests that many Filipino employees may already be more AI-comfortable than their managers realize, because personal use is happening outside of formal training. Business leaders should explore what their teams already use AI for personally, and build internal adoption programs that bridge those existing habits into work contexts rather than starting from zero.
How can business leaders encourage AI adoption without mandating it?
Start with visibility: use AI tools openly in meetings, share what you tried and what worked. Create low-stakes opportunities for teams to experiment. Ask questions rather than issuing policy. People adopt what they see working around them before they adopt what they’re told to use.
Is it safe for employees to use AI for personal matters at work?
It depends on the tools and the data involved. Employees should avoid entering company-confidential information into personal AI accounts. Organizations need clear, practical policies that address this, not blanket bans that push behavior underground. A good AI governance policy distinguishes between personal-use habits and data risk, rather than treating them as the same issue.
