She Built a Product That Reached 3 Million Users. Now She’s Building AI She Hopes You’ll Need Less.

AI can increasingly help people decide what to do and how to do it. For Lan Doan, that makes another question more important: why are we doing it at all?

Lan is the founder of Asla, whose newest product, Ren, is built around a first-principles bet: as AI takes on more of the “what” and “how” of our lives, the ability to understand our own motivations, interpretations, and relationships becomes increasingly important.

A former BCG consultant, Harvard MBA, and former CEO of a product that reached three million users, Lan is now completing an MSc in Psychology & Neuroscience of Mental Health at King’s College London. She is building Asla from Ho Chi Minh City, Vietnam.

Rather than positioning Ren as another AI coach, Lan is exploring how technology can help people strengthen something she believes should remain their own: the ability to understand themselves, question their interpretations, and connect with others.

AI Can Give You the Answer. Lan Wants You to Know Why You Asked. 

Lan sees this moment in AI as an opportunity to reconsider what people can delegate to technology.

“You can rely on technology to own the what and the how — and you should,” she says. “That frees us up to get really good at the one thing that has to stay ours: the why.”

For Lan, the “why” is both practical and profound. Someone who is clear about their motivations can decide faster, waste less, and avoid being pulled around by every confident opinion in the room.

But it is also about agency.

“Owning your why is how you own your agency — and your agency is most of what your life actually is. It’s where meaning lives.”

For a long time, she says, people were so busy with the “what” and “how” that they rarely had space to examine why they did — or did not do — certain things. AI could create that space.

The risk is becoming dependent on it before developing a sufficiently strong “why” of our own.

Escaping the Automated Echo Chamber

A conversation with a friend made that risk particularly vivid for Lan.

Her friend spends hours talking to AI and excitedly told Lan that she had discovered something: the AI agreed with her long-held belief that metacognition is what separates humans from machines.

Lan asked her own version of the question.

The AI told her that the difference was knowing your “why” — exactly what Lan already believed.

“It agreed with both of us.”

For Lan, the concern was what could happen next.

“It had quietly built each of us our own echo chamber. And the risk there isn’t the agreement itself — it’s that once something feels settled, we stop checking with other people, stop discussing it, and stop letting anyone else complicate the picture.”

Ren is designed around that tension. It takes what happened to a user and helps them examine it from different perspectives. Rather than telling them what to think, it encourages them to check their interpretation, consider why they formed it, develop metacognition, and remain curious about other people.

It also connects users to anonymous chapters from others, allowing them to see aspects of themselves through someone else’s experience.

For Lan, being able to answer “Who am I?”, “Why do I do what I do?”, and “What role do I want to play in my community?” is part of owning the “why.”

What If AI Helped You Connect — Then Got Out of the Way?

Lan connects this philosophy to her MBA training, where managing oneself and connecting with others were part of becoming a better leader.

“Each of us is now a mini-CEO of an army of agents — and a CEO should know their why very clearly.”

It is also why she deliberately places Ren outside the AI-coaching space.

General-purpose AI can already ask questions, remember context, personalise responses, and maintain conversations. Lan does not want Ren to compete by becoming a better artificial relationship.

“Most AI coaching is trying to solve problems by providing better answers, holding more context, building a relationship with the user, and so on. AI coaches aim to replace human connection.”

Ren focuses instead on building the skill to connect.

“We don’t solve loneliness by offering a replacement for human connection; we solve it by raising people’s capacity to connect.”

That usually begins, she says, with understanding oneself and becoming more open to seeing others.

Lan also sees this distinction as part of Asla’s defensibility. She is not betting on having a cleverer model, but on what she sees as a durable human need.

The more people understand themselves and others, she argues, the more deeply they want to know them. If Asla continues in that direction, she sees opportunities to build more dimensions of connection, support deeper connection within communities, and potentially enable the pairing of connected people.

Three Million Users Taught Her That Attention Isn’t Impact 

Lan has scaled a product before, but she is careful about claiming that experience gives her all the answers for Asla.

“Honestly, it’s early for me to talk about scaling with any authority, and I’m deliberately staying flexible on how we get there.”

What has changed is the environment in which products are built.

“Last time, building the system was the hard part, and doing it well was the moat. That world is gone.”

Today, she argues, almost anyone can ship a product. Distribution has become the harder challenge.

That changes the allocation of effort: less emphasis on “Can we build it?” and much more on “Can we reach the people who need it and earn their trust once we do?”

One principle carries over from her previous experience: understand stakeholders more deeply than anyone else bothers to, then be inventive about the system built around them.

But three million users also taught Lan to distinguish attention from impact.

“Three million users, if I’m honest, mostly taught me what people will click — not whether I changed a single life.”

This time, she says, she would rather be “irreplaceable to a few than glanced at by millions, and grow from there.”

There are also two things she refuses to sacrifice for that growth.

Ren Refuses to Give You the One Thing Everyone Wants: An Answer 

The first is a verdict.

Ren will not tell a user, “You were right.” It will not say, “Leave him.” Nor will it hand them a five-step plan.

“Verdicts are the most addictive product on the internet; whole forums run on strangers voting on someone’s life.”

Lan knows that restraint can come at a cost to engagement. But giving users the answer would undermine what she wants Ren to encourage.

“A verdict ends the curiosity, and the curiosity is the point.”

“The day we start answering for you, we’ve become the thing we set out to improve on.”

The second principle concerns ownership.

Your Most Personal Data Should Never Become the Product’s Leverage 

Users’ original words are kept exactly as they wrote them. Ren’s narration is intended as a lens, never a replacement.

Sharing is de-identified. Users can leave whenever they choose, taking everything with them or erasing everything completely.

“We hold people’s most private material; the only acceptable posture is custodian, never owner.”

For Lan, no growth tactic is worth breaking that principle.

Her approach therefore puts limits on some of the mechanisms that could make a deeply personal product more engaging or difficult to leave. Ren will not create certainty by handing down verdicts, and the personal material users accumulate remains theirs.

That leaves Asla with a different question: if time spent in the product is not the ultimate objective, what does success look like?

Her Measure of Success? You Close the App. 

Lan has a straightforward answer.

“If Asla works, the proof won’t be hours logged in an app.”

It will be the people who understood themselves, understood someone else, and walked back into their own lives — and then perhaps sent a friend to the door.

That brings the product back to the idea it began with.

As AI becomes increasingly capable of handling the “what” and “how,” Lan wants people to retain ownership of the “why”: their motivations, interpretations, relationships, and ultimately the choices they make.

Ren can help someone look deeper, encounter another perspective, and question an interpretation. But it is deliberately not designed to provide the final verdict.

That creates an unusual definition of success for an AI product.

If Asla works as Lan intends, its value may not be demonstrated by how long someone stays inside Ren.

It may be demonstrated by what they take with them when they leave.

Read the Chinese article here.