The Lawyer Who Built an AI So Abuse Survivors Don't Have to Investigate Their Own Cases

From Jakarta to Stanford

Puteri describes growing up with little money and a clear ambition: to help her parents escape poverty. She became a family breadwinner as a teenager. Debate competitions offered prize money and helped open a path toward law school at the University of Indonesia.

Her early jobs taught her skills she would use later. Selling food meant approaching strangers and handling rejection. Tutoring taught patience; journalism taught her to listen and ask better questions. She subsequently worked in legal and technology roles in Indonesia, including at Dealls and GovTech Edu, while pursuing an education-focused community initiative.

She eventually completed an LL.M. at Stanford Law School, where she says she studied feminist and gender issues and received recognition for pro bono work. Starting from zero, she says, was already familiar.

Knowing the law is not the same as being able to use it

At 17, Puteri says she experienced what she later recognized as coercive control and financial exploitation in a personal relationship. She describes being pressured into using money from a student organization she led, then bearing the consequences. For years, she lacked the language to describe what happened.

“I thought abuse had to involve physical violence. I didn't know that someone could manipulate your emotions, exploit your trust, and gradually take control of your decisions without ever laying a hand on you.”

She also recounts experiencing sexual harassment in 2023, followed by threats to her reputation, and says she later faced online harassment while building her company. These are her accounts of events involving others; LEVEL has not independently established those allegations.

What she describes in detail is the work that followed: collecting screenshots, preserving messages, identifying dates, reconstructing conversations and repeating painful accounts to different professionals. At times, she says, rereading messages felt unbearable.

“If this was so difficult for me, what happens to someone who doesn't have my education, resources, or support network?” she asked herself. Puteri began teaching herself to code and experimenting with AI prototypes before assembling a team to build Lea.

What Lea is designed to do—and what it cannot

According to Puteri, Lea helps users describe concerning experiences in their own words, identify potential patterns of coercion and organize messages, emails, photographs and other records into a chronology. A user might ask whether a partner's phone monitoring or threats are warning signs; the app is intended to explain potential concerns and options without making decisions for them.

Puteri says the system is designed to recognize more than 130 patterns of abuse. The company describes a protected evidence vault, PIN access, Quick Exit and Disguise Mode. Other integrations remain in development. LEVEL has not independently tested the app's detection accuracy, security architecture or evidentiary outputs.

Those distinctions matter. An AI tool handling intimate accounts and potentially dangerous evidence needs unusually strong safeguards. A product description is not proof of privacy or reliability.

Puteri acknowledges that Lea is not a lawyer, therapist or emergency service. It cannot determine whether a crime occurred or guarantee that a record will be admissible in court. AI can make mistakes; human review remains essential.

The myth of the perfect survivor

One story Puteri remembers involves a domestic-violence attorney who needed to hire another lawyer to obtain her own protection order. Professional expertise did not remove the practical and emotional difficulty of seeking help.

Puteri says she has spoken with more than 1,000 survivors while developing Lea. The conversations challenged a common expectation: that someone experiencing abuse should leave immediately, document everything perfectly, remember every detail and remain calm throughout.

“We tend to look for a perfect victim, someone who behaved exactly as we think a victim should. But real people are complicated, and trauma is complicated.”

Financial dependence, fear, affection and concern for children can complicate decisions. Seeking help can require telling the same story repeatedly. AI might reduce some of that administrative burden, but it cannot replace institutions willing to act, qualified professionals or meaningful protection.

Who pays for safety?

Puteri says Legali AI is developing a model in which universities, employers, shelters and legal-aid organizations pay to provide Lea to people they serve. The premise is that people seeking help should not necessarily bear the cost themselves.

Whether institutions will fund it, how access will work and how sensitive information will be protected remain consequential questions. The company's promises must ultimately be measured against demonstrable safeguards and outcomes.

“I don't want people to need Lea forever,” Puteri says. Her ambition is to help them regain control of their lives without spending years reconstructing and defending what happened.

Her journey from teenage work in Jakarta to building technology in Silicon Valley leaves a question larger than any app: Why does getting protection still demand so much work from the people who need it?

Reporting note: This profile draws on written answers Ira Puteri provided directly to Revista LEVEL. Product claims and sensitive allegations are attributed to her and have not been independently verified by LEVEL.