KINNECT: Using AI to Build a Private, Secure Platform for Family Memories

"They responded well to feedback and ideas within an appropriate response timeframe. They're great. "
- Omar Alvarez, Founder & CEO Kinnect

About the app

Did you know that loneliness is as harmful as smoking 15 cigarettes a day?

KINNECT is a mobile app designed to reduce loneliness by creating deeper connections among families through a private, digital space where you can capture, share, and relive memories.

At its core, KINNECT allows users to create "StoryBooks" where they can document personal experiences through audio, video, and written formats.

The app’s main goal is to strengthen relationships, reduce loneliness, and preserve family legacies by turning everyday moments into lasting memories.

Developed with an emphasis on ease of use, inclusivity, and emotional connection, KINNECT addresses the rising issue of social isolation, particularly among seniors, while creating a user-friendly environment for all ages.

The platform integrates advanced AI technologies for personalized prompts, memory categorization, and emotional analysis to enhance the storytelling experience, which is exactly what our goal was in this project.

Screenshots from Kinnect's website home page with 3 screens from the app and each has a title: "For your Kin of today and in the future", "Pre-created books save time and effort" and "Feel all the feels, anytime"

Source: Kinnect

Challenges

Screenshot from Kinnect's About web page. On the right there are 2 screenshots of the app and on the left it says: Kinnect is for everyone. Inclusive and diverse, deeply personal and simple yet meaningful.

Source: Kinnect

1. Personalized Storytelling Prompts

The primary challenge was designing an AI system that could generate personalized prompts based on user data, preferences, and past interactions. The AI needed to evolve with user input, learning from their responses to suggest more relevant themes for storytelling.

2. Emotion Recognition & Sentiment Analysis

Understanding the emotional tone of the user-generated content was important to ensuring the app could foster deeper, meaningful interactions. The AI had to interpret emotions from both written and multimedia content, adjusting prompts based on the user’s emotional state or the tone of the stories they shared.

3. Categorizing and Organizing Memories

With diverse types of memories—ranging from family milestones to personal reflections—the AI needed to automatically categorize and organize these stories. This included identifying themes, relationships, and emotions from user data to make memories easier to browse and recall.

Solutions

Screenshot from Kinnect's How it Works page with 3 screenshots of the mobile app on the left and explanation for each on the left: Easy on purpose. 1 Create a KIN group, 2. Capture your stories your way, 3. Feel all the feels, anytime

Source: Kinnect

The solutions found helped us deliver a successful product using AI Ideation, AI Build and AI Delivery.

1. AI-Driven Personalized Prompts

We developed an AI system that generates prompts tailored to each user's preferences, past stories, and interactions with family members. By representing the information using Knowledge Graphs, which allow easier interrogations to be performed using a specific query language, we managed to identify easier key points and specific details and generate new meaningful prompts.

2. Emotion Recognition & Sentiment Analysis

Applying sentiment analysis techniques, we are able to detect things like specific emotions, moods or changes of mood. This allows the AI to suggest prompts that align with the user's current mood, helping to foster deeper emotional engagement. The system can detect happiness, sadness, nostalgia, and more, adjusting the content to create a more empathetic experience.

3. Memory Categorization & Organization

We implemented an AI engine that categorizes memories based on themes like "Family Milestones" or "Life Lessons." By identifying key elements such as people, emotions, and activities, the AI organizes stories into meaningful categories, making it easier for users to search, browse, and relive their memories.

Four screens of the app from the Apple Store.

Source: App Store

To be able to complete the project, we used Natural Language Processing (NLP) for analyzing user-generated text and detecting emotional tone and key themes, Machine Learning Algorithms for dynamic personalization of prompts and adaptive learning from user interactions, and Graph Analytics to identify connections between family members and shared memories for prompt generation.

To analyze audio and video content for emotional recognition and memory categorization we used Multimedia Processing and then integrated Sentiment Analysis into the AI to gauge emotional context and adjust prompts accordingly.

Want to work with us? Send us a message at hello@x2mobile.net or fill out this form. We reply in 1 business day!