MeMind and Yucatán used tailored AI screening to reach 10,000 at-risk people

Yucatán faced a suicide rate twice Mexico's national rate and a conservative culture in which families often hid the cause of death and many people were reluctant to seek care. Connectivity and provider access were also uneven.

In 2022, the state government partnered with MeMind, whose app asks a 15-minute mental-health questionnaire and uses an AI diagnostic tool to assign risk categories. The Spanish version was tailored with the state's Department of Health around domestic violence, alcoholism, and substance abuse; severe-risk users received periodic check-ins, and consenting users' data could be shared with government.

Two months after launch, the governor said more than 200 high-risk cases had been identified; the government later attributed a 9% statewide suicide-rate reduction in 2023 to the app. Over two years, MeMind deemed 10,000 users at risk and gave them treatment information; total users reached 80,000, with 50% still using the app after six months. The limits remained visible: the app needed log-ins, young men used it least consistently, and many residents lacked phones.

Yucatán reported a 9% suicide-rate reduction in 2023; MeMind reached 80,000 users and connected 10,000 at-risk people to treatment information over two years.

The state's suicide rate and reluctance to seek care required a low-friction screening route.

Localizing questions around domestic violence, alcoholism, and substance abuse made the assessment more relevant than a generic survey.

Risk categories triggered periodic check-ins and treatment information, turning detection into a referral path.

Government access to consented population data let the program inform public-health policy.

Tailor an AI screening tool to local risk patterns, then pair risk detection with human treatment pathways; a score alone does not treat a patient.

MeMind planned to add chatbot therapy, but the article also records the constraints: coverage depended on users completing check-ins; young men—the group most likely to die by suicide—were least consistent; and residents cited economic stagnation and limited phone access. The model therefore extended screening reach without removing the need for clinicians, infrastructure, or broader social intervention.