Mirra

Mirra is an AI-powered chatbot that helps parents bridge this gap through a science-based reflection process

What is it?


Many parents experience frustration when communicating with their teenage children. Mirra is an AI-powered chatbot that helps parents bridge this gap through a science-based reflection process. In this process, the parent observes and delves into the interactions and arguments they have had with their teenager. Reflection has been shown to help increase awareness, deepen understanding of the relationship, identify behavioral and parenting patterns, and identify new ways to cope with the child. Thus, using the chatbot will lead the parent to new insights, effective coping mechanisms, and improved communication with their teenager.
As parents engage with the chat, Mirra continuously learns and adapts to their unique needs and challenges. This ensures that the parent receives personalized and accurate suggestions for their specific situation, leading to more effective communication and stronger relationships with their teenagers

Project Info


Students: Ariel Cohen, Naama Israel, Ron Rozen, Omri Aboudi, Ori Grotto
Year: 2024
Lecturer: Dr Noa Morag
Teaching Assistant: Ofir Sadka

Students' Website

How does it work?

After downloading the app the parent goes through a quick signup process so that the chat can get to know the parent, the teenager and the challenges in their communication. For example, the app asks: “What is your teenager’s biggest passion?” or “What is the biggest challenge when communicating with your teenager?”. Once this initial setup is complete, a chat session begins, initiating the reflection process.
The parent starts by sharing a recent interaction with their teenager that was particularly emotional or frustrating. The chatbot then asks reflective questions like, “How often do situations like these occur?” or “Is there another way you think you could’ve responded to your child’s behavior?” These questions are based on a scientifically proven reflection protocol taught to a GPT-4 model.
At the end of the conversation, the app encourages the parent to write down their insights to help internalize them. We want the parent to actively translate the interaction into clear insights- a process that is known to enhance learning and behavior change. After writing the insights, the parent is offered advice on different approaches he can try in similar situations in the future. The suggestions given by the chatbot are grounded in expert parental guidance and scientific research.
Finally, all the user’s insights, reflections, and advice are saved within the app, allowing parents quick access as a reminder of what they have learned.