Case Study

Designing a Smarter
AI Companion

Jaimee AI’s Approach to Memory, Sentiment, and Women’s Emotional Health

woman graphicman graphic
Introduction

Jaimee AI, an AI companion, is addressing a critical gap in emotional wellness for adult women, particularly frontline workers in high-stress, service-oriented roles like childcare, nursing, cleaning, and hospitality. These individuals often shoulder immense emotional burdens while navigating the dual pressures of professional and personal care responsibilities, often without access to meaningful support. Jaimee AI provides a safe, non-judgmental digital space for connection and validation, where women can vent, talk through challenges or simply enjoy some playful banter to help escape the daily stress.

Jaimee AI is not only a tool for emotional wellness but also a response to the gender bias deeply embedded in AI systems. The platform was founded on the recognition that many AI models are trained predominantly on male-centric datasets, leading to responses that often fail to reflect women's lived experiences. Jaimee AI challenges this imbalance by prioritizing women-centered data to inform its interactions; this began with modeling the communication style on that of the female developers and is being extended into identifying male bias or sexist language in the chat and removing it from the training data. The goal is twofold: to create an AI companion that genuinely resonates with female users and to contribute to the broader movement of making AI more inclusive. Behind this vision are co-founders Sreyna Rath (CEO) and Camilla Cooke (CMO), who have championed a mission to build AI that serves women authentically and equitably. By fostering an alternative to hyper-sexualized or one-dimensional portrayals of women in AI, Jaimee AI ensures that its users—whether seeking friendship, mentorship, or romance —engage with a system designed with their needs in mind.

Behind this groundbreaking tool is Zep AI, whose memory, sentiment analysis, and security capabilities enable Jaimee AI to achieve its mission while meeting the high technical standards required to build user trust and engagement.

jaimee device
Technical Problem Statement

Jamiee AI faced three core technical challenges in its development

Incorporating a Concept of Time: AI systems inherently lack an understanding of time, a critical feature for users seeking follow-ups or reminders. Jamiee AI needed a mechanism to remember and act on user interactions at future moments.

Sentiment Analysis and Tailored Responses: While emotional resonance is central to Jaimee, the platform struggled to interpret user sentiment effectively and tailor appropriate responses. This capability is vital to ensuring users feel engaged, validated, and supported.

Data Security and Privacy Compliance: Handling sensitive emotional conversations requires robust privacy protections and adherence to regulatory standards, such as SOC 2, to ensure user trust and scalability in global markets.

solution device
Solution

Building the Right Companion with Zep AI

To address these challenges, Jaimee AI partnered with Zep AI to enhance prompting, functionality, and user experience. During the beta phase, the team piloted the platform with numerous users and iterated on engagement metrics and UX design throughout their time at Techstars Sydney Accelerator program. This iterative process highlighted the need for more emotional resonance and personalized follow-ups, leading to enhancements in sentiment analysis and context retention. By prioritizing a user-driven development approach, Jaimee AI continues to evolve in alignment with the diverse needs of its audience, from tech-savvy early adopters to mainstream frontline workers.

Efficient Knowledge Retention:

Zep AI’s memory system allows Jaimee AI to retain and organize conversational context, ensuring meaningful follow-ups. This capability addresses the lack of a native "time" concept in AI, enabling Jaimee AI to remember interactions and re-engage with users effectively

Enterprise-Grade Security:

Zep AI’s SOC 2-compliant systems provide a secure environment for Jaimee AI, safeguarding user data while adhering to stringent regulatory standards. This is crucial for a platform handling sensitive conversations and aiming for large-scale adoption in diverse markets.

Looking ahead

Jaimee AI recently launched on the app store, marking a significant milestone in its mission to improve emotional wellness. With an oversubscribed waitlist and upcoming corporate pilots, Jaimee AI will continue integrating user insights to refine and expand its offerings. Future developments include:

Advanced Graph Features:

Zep AI’s graph capabilities will enable deeper user insights and more nuanced sentiment analysis, supporting more personalized and impactful interactions.

Proprietary Model Training:

Jaimee AI plans to move beyond OpenAI by training a proprietary LLM, leveraging Zep AI’s infrastructure to customize functionality and build bias-free, inclusive guardrails.

By harnessing Zep AI’s technology, Jaimee AI is uniquely positioned to lead in the emotional wellness space, offering a scalable, secure, and impactful solution tailored to the needs of women that will engage more women with AI, which in turn will help reduce male bias overall. Together, the partnership exemplifies how technical excellence and a mission-driven approach can combine to create transformative tools for underserved communities.