Key Takeaway
AI becomes more interesting to me when it moves beyond generating outputs and starts solving real business problems and improving workflows.
AI became interesting to me when it stopped feeling like something happening somewhere else.
I started using it.
Not because I had a complete understanding of artificial intelligence, but because I started seeing practical problems where it could help.
Creative work.
Product photography.
Content.
Video.
Research.
Workflows.
And eventually, business processes.
That changed the question I was asking.
Instead of:
**"What can AI generate?"**
I started becoming more interested in:
**"What can AI actually help a business do?"**
I'm still exploring that question.
## I started with creative work
Some of my earliest practical experiences with AI came through creative work.
I've used AI-assisted approaches for product photography, creative development and visual content.
For TOHVA, for example, I worked on AI-powered product photography as part of the broader content-production process.
I also used AI-assisted visual development while working on the Pooja Vysh Makeovers project, including the brochure.
These experiences showed me something simple:
AI doesn't have to be a completely autonomous system to be useful.
Sometimes it can simply make one part of a workflow faster or more flexible.
A person still has to understand the requirement.
The output still needs to be evaluated.
The result still needs to fit the brand and the objective.
AI is a tool inside the process.
That distinction has stayed with me.
## Then I started looking beyond the output
Creative generation is probably one of the easiest ways to notice what AI can do.
You give it something.
It produces something.
You iterate.
You improve it.
But businesses aren't made of outputs alone.
They are made of processes.
A piece of content is one output.
Behind that content might be:
Research.
Ideation.
Briefing.
Creation.
Review.
Editing.
Approval.
Publishing.
Measurement.
Iteration.
That made me more interested in what happens when AI moves from generating an individual output to becoming part of the workflow itself.
## The question became: what problem are we solving?
I don't think AI should be added to a business simply because AI exists.
The better question is:
**What problem are we trying to solve?**
Maybe a process is repetitive.
Maybe information needs to move between systems.
Maybe a team spends too much time doing the same manual task.
Maybe research takes too long.
Maybe content production has too many repetitive steps.
Maybe a business has information but isn't using it effectively.
Those are more interesting AI opportunities to me than simply asking which tool is currently popular.
The technology is interesting.
But the problem comes first.
## AI and marketing naturally came together for me
Marketing was already a major part of my work before AI became such an important part of my thinking.
That made the intersection particularly interesting.
Marketing contains a lot of activities that involve research, writing, analysis, creative production, testing and iteration.
AI can assist with many of these.
It can help generate ideas.
It can help explore different approaches.
It can assist with research.
It can accelerate certain creative tasks.
It can help structure information.
But there is an important limitation.
AI can generate possibilities.
It doesn't automatically understand what the business should do.
Understanding the customer, the positioning, the offer and the business objective still matters.
That's why I see AI as an amplifier of a process rather than a replacement for thinking about the process.
## From AI outputs to workflows
This is where my interest started moving toward automation.
Using AI once is one thing.
Building a repeatable workflow around it is another.
For example:
**Input**
↓
**Process**
↓
**AI**
↓
**Review**
↓
**Output**
↓
**Measurement**
↓
**Improvement**
The interesting part isn't necessarily the AI model in the middle.
It is the system around it.
What information goes in?
What happens next?
Where does human judgment remain necessary?
What can be automated?
What needs approval?
How do we know whether the result is actually useful?
These questions have made me increasingly interested in systems and automation.
## I'm exploring AI agents too
Another area I'm currently exploring is AI agents.
The idea is interesting because it moves beyond asking AI to perform one isolated task.
Instead, an AI system could potentially work through multiple steps toward an objective.
But I'm approaching this as a learner.
I'm not presenting myself as an AI-agent expert.
I'm trying to understand:
What should an agent actually do?
Where does an agent make sense?
What information should it have access to?
Where should humans remain involved?
What happens when it makes a mistake?
How can these systems be made reliable enough to be useful?
Those questions matter more to me than simply building something that looks impressive in a demo.
## AI needs business thinking around it
One of the things I keep coming back to is that AI doesn't remove the need for business thinking.
If anything, it can make that thinking more important.
If everyone can generate content faster, differentiation may move toward the quality of the idea.
If everyone can automate certain tasks, the important question becomes which tasks are worth automating.
If everyone can access powerful AI tools, the advantage may increasingly come from how those tools are connected to a company's data, workflows, people and customers.
That's why I don't see my interest in AI as separate from my interest in business.
They are becoming increasingly connected.
## My current AI × Business framework
Right now, I'm thinking about AI applications roughly like this:
**Problem**
↓
**Process**
↓
**AI Opportunity**
↓
**Experiment**
↓
**Evaluate**
↓
**Improve**
↓
**System**
The important word for me is **experiment**.
I don't think every AI idea needs to become a product.
Sometimes you try something and discover that it isn't useful.
Sometimes the technology works but the workflow doesn't.
Sometimes the problem isn't important enough.
Sometimes a simpler solution is better.
That learning is part of the process.
## I'm still exploring
I'm still early in this journey.
I've used AI in practical creative work.
I've explored automation.
I'm learning about agents and newer AI systems.
I'm interested in how AI can connect with marketing, business processes and digital products.
But I don't have a final answer for what AI will mean for my own work.
And I don't think I need one yet.
What interests me is experimenting.
Taking a real problem.
Finding out whether AI can help.
Building something.
Testing it.
Learning from what happens.
Then improving the system.
That's increasingly how I want to approach AI.
Not as something to blindly follow.
Not as something to fear.
And not simply as another tool to collect.
**I want to understand what can actually be built with it.**
And more importantly:
**what is actually worth building with it.**