Her 11-Year-Old Built a Dog App. Swetha St. Peter Says the Future of AI Is Agents.

By Li Lin | Legacy Launchers

A lot of people are still trying to figure out how to use ChatGPT.

Swetha St. Peter is already thinking about what comes next.

On this episode of Legacy Launchers, I sat down with Swetha, an AI and product expert with experience across organizations including Accenture, Amgen, Disney, and MIT Sloan, to talk about agentic AI, empathy-driven design, the future of work, and one of my favorite examples of AI in action:

Her 11-year-old daughter built an app for their Maltese.

That one story says a lot about where AI is going.

The next generation is not waiting for a perfect roadmap. They are building useful things around real problems.

And that is exactly how Swetha thinks businesses should approach AI too.

We are moving from AI that answers to AI that acts

For the last few years, most people have experienced AI through tools that generate text, summarize information, brainstorm ideas, or answer questions.

Swetha sees the next major shift as agentic AI.

Instead of simply telling you what to do, an AI agent can potentially complete parts of a workflow.

She gives a healthcare example.

A traditional chatbot might explain how to reschedule an appointment.

A more capable assistant might help someone find available times.

An AI agent, with the right permissions and safeguards, could potentially verify the patient, understand the request, check availability, book the appointment, update the relevant systems, and send confirmation.

That is a very different level of usefulness.

The distinction is simple:

Generative AI gives you an answer.
Agentic AI takes action.

For business owners, that means the AI conversation is moving beyond “Can this write my email?” toward “Can this actually help complete the process?”

Stop asking where you can add AI

One of the strongest ideas in the conversation is also one of the simplest.

Swetha says companies are starting to realize that adding AI everywhere does not automatically create value.

The better questions are:

What problem are we solving?
What outcome are we trying to create?
Should AI even be used here?

That last question matters.

There is a temptation right now to force AI into every workflow because it sounds innovative.

But technology is only useful when it improves something.

Faster does not automatically mean better.

Automated does not automatically mean smarter.

And more AI does not automatically mean more value.

The goal is not to have the most AI.

The goal is to solve the right problem.

Empathy should come before the product

Swetha’s work in product management and design thinking has made empathy a major part of how she evaluates technology.

For her, empathy is not something you add at the end.

It starts before the product is built.

Who is going to use this?

Who might be excluded?

What happens if the model gets something wrong?

Are you designing for a real range of users or only for an imagined “average” person?

Those questions become especially important when AI starts making or influencing decisions that affect people.

Technology is trained on human-created data.

Human data contains human patterns, assumptions, and biases.

If those patterns are not examined, AI can reproduce them at scale.

That is why human judgment still matters.

AI can sound confident and still be wrong

During our conversation, I shared an experience where I asked AI to help price my services and felt like the suggested number undervalued me.

When I questioned it, the system gave me an explanation that referenced gender and race.

Swetha pointed out something important.

Just because AI gives a confident explanation does not necessarily mean that explanation accurately describes what happened internally.

That is an important distinction.

AI can produce a convincing story about its own answer without truly “knowing” why that answer happened in the way a human might assume.

The lesson for business owners is bigger than pricing.

Do not confuse confidence with truth.

Use AI as a tool.

Check the assumptions.

Compare the output against real-world evidence.

And do not outsource your judgment to a machine.

Her 11-year-old daughter is already building

Then we got to one of my favorite parts of the episode.

Swetha’s daughter is 11.

She wanted to build an app related to their Maltese.

The idea was to track things like the dog’s training habits and eating habits.

Instead of giving her daughter a giant technical lecture, Swetha asked simple product questions:

Why do you want to build this?

What is the value?

What problem are you solving?

Once the purpose was clear, the milestones became easier.

Her daughter completed the first phase as a weekend project.

That is wild when you think about it.

A lot of adults are still debating whether they should “learn AI.”

An 11-year-old is already using the tools to build around a real need.

And Swetha is teaching her the more important skill:

Not just how to build.

How to think.

The kids may have an advantage

There is something interesting happening with younger people and AI.

Many of them are not carrying the same intimidation around technology.

They are not asking whether they have permission to build.

They are experimenting.

Swetha said her daughter constantly comes up with ideas, so she gave her a notebook and encouraged her to write them down.

That combination is powerful:

Ideas.

Curiosity.

A place to capture them.

And tools that can help turn those ideas into something real much faster than previous generations could.

The competitive advantage may not belong to the person who knows the most technical jargon.

It may belong to the person who can identify a useful problem and move quickly enough to test a solution.

The future of AI is quieter than people think

Swetha believes AI is moving toward becoming less of a separate tool and more of an intelligent layer around everyday life.

That might actually be the biggest shift.

Eventually, using AI may feel less like opening a special application and more like using electricity or the internet.

You stop thinking about the infrastructure.

You just expect things to work.

That is where agents become especially important.

Instead of constantly prompting a tool, people may increasingly delegate outcomes.

Book this.

Research that.

Update this.

Coordinate these systems.

Handle this workflow.

Then escalate to a human when needed.

The interface starts to disappear.

The result becomes the product.

AI is a double-edged sword

Swetha does not pretend AI is automatically good.

She describes it as a double-edged sword.

It can remove repetitive work.

It can improve decision-making.

It can accelerate research.

It can support scientific discovery.

It can personalize education and healthcare.

But the value depends on how it is used.

That is why she keeps returning to intention.

Technology does not remove the need for good judgment.

It raises the stakes for it.

The more powerful the tool becomes, the more important it is to ask who is directing it and toward what outcome.

Don’t fear what you have not tried to understand

Toward the end of the conversation, I asked Swetha what she would say to people who are convinced AI is going to destroy everything.

Her answer was simple:

“Don’t fear what you haven’t taken the time to understand.”

That does not mean ignoring legitimate risks.

It means curiosity is usually more productive than panic.

Every major technological change creates fear because it disrupts familiar ways of working.

AI is no different.

The productive move is to learn enough to make informed decisions.

Start small.

Pick one real problem.

Test something.

See whether it saves time, improves quality, or makes a process easier.

Then build from there.

The future of AI is agents

The biggest takeaway from Swetha’s episode is not that everyone needs to become a programmer.

It is almost the opposite.

The future may belong to people who understand problems well enough to direct increasingly capable tools toward useful outcomes.

That is what good product managers already do.

It is what good business owners do.

And apparently, it is what some 11-year-olds are already doing too.

The next era of AI is not just about generating more content.

It is about getting useful work done.

The future of AI is agents.

And the people who learn how to guide them well may have a very big advantage.

🎙️ Listen to Swetha St. Peter on Legacy Launchers with Li Lin for the full conversation on AI agents, empathy-driven design, product thinking, the future of work, and what the next generation is already building.