BiteWise – Relearn Eating; A Judgment-Free AI Eating Coach

BiteWise is a judgment-free AI coach that helps people relearn conscious eating and build lifelong healthy habits. Instead of prescribing what to eat, it reverses the approach: users share how they currently eat, and the AI helps them understand how their choices impact their body, and how small, gradual changes can significantly improve their health without giving up the joy of eating. It offers real-time support throughout the day. Inspired by glucose-stability principles and behavioral nutrition.

don't have one yet

Giulia Jagan

giuliajagan@gmail.com

9230 Flawil

The core problem is not a lack of knowledge about healthy eating, but a structural gap between knowledge and real-life behavior. Existing solutions rely on prescriptive diets, static plans, or calorie tracking, assuming that rational information leads to better decisions. In reality, eating decisions are emotional, contextual, and made in the moment—often under stress, fatigue, or cravings. This creates a systemic failure: people are educated on what is “right,” but unsupported when it actually matters. As a result, they experience guilt, inconsistency, and unsustainable habits, eventually disengaging from healthy eating altogether. BiteWise addresses this behavioral gap by shifting from instruction to understanding. Instead of telling users what to do, it helps them interpret their own habits, understand the impact of their choices, and learn how small, manageable adjustments to how and what they eat can significantly improve their health. This builds awareness and motivation in real time, enabling users to make better decisions in the moment and supporting lasting behavior change rather than short-term compliance. This creates a systemic failure: people are educated on what is “right,” but unsupported when it actually matters. As a result, they experience guilt, inconsistency, and unsustainable habits.

The goal is to replace reactive, emotionally driven eating with conscious decisions that work under real-life conditions. Today, eating habits are shaped by a combination of constraints: people are tired, in a rush, or emotional—but also limited by budget, time, and the availability of options around them. In these conditions, decisions default to what is easiest, fastest, most affordable, and most comforting. This leads to reactive, inconsistent eating patterns— sometimes even followed by frustration or guilt. The problem is not a lack of knowledge. It’s that better options don’t win at the moment of choice. BiteWise focuses on this exact moment. Instead of prescribing ideal diets or adding more information, it supports users in real time—when decisions are actually made. Through simple, personalized guidance, it helps people navigate trade-offs and find better options that still fit their situation. The approach is based on small, achievable adjustments rather than radical change. By making better choices feel doable and worthwhile within real-life constraints, these decisions can be repeated, reinforced, and gradually turned into habits. Over time, this shifts behavior from reactive and inconsistent to more intentional and sustainable—without requiring perfection or drastic lifestyle changes.

BiteWise supports individuals at the exact moment where food decisions are made—when intention meets real-life constraints. People don’t struggle because they lack knowledge. They struggle because, in everyday conditions, better options don’t win. When people are tired, in a rush, emotional, or on a limited budget, they choose what is easiest, fastest, cheapest, and most comforting. In these moments, “good food” consistently loses—not because it’s undesirable in theory, but because it doesn’t feel doable in practice. Despite the existence of nutrition guidelines, health campaigns, and local food initiatives, there is a critical gap: none of these operate at the moment of decision. BiteWise addresses this gap directly. It acts as a real-time companion, helping users make better choices feel easy, desirable, and worth it—by offering practical, context-aware alternatives rather than ideal but unrealistic solutions. The focus is not on perfection, but on small, achievable improvements that fit into real life. This is where impact scales. Small, repeated decisions shape habits. Habits shape demand. And demand influences what is produced, offered, and normalized. By increasing the likelihood that better choices win in real-life situations, BiteWise contributes to a broader shift in how food is consumed and valued. Over time, this can support improved population health, reduce pressure on healthcare systems, and contribute to reshaping food environments—making healthier options not just available, but naturally preferred.

The idea has been informally tested through daily use of AI tools to support real-time eating decisions. This included asking for guidance during cravings, food selection, and supplement timing. The key learning is that immediate, contextual support is significantly more effective than static advice. Users do not need more information, but support at the moment decisions are made. Additionally, a non-judgmental tone and flexible guidance (e.g. allowing compromises rather than restriction) proved critical for maintaining engagement and reducing guilt-driven behaviors. As well as Additionally, a non-judgmental tone and flexible guidance (e.g. allowing compromises rather than restriction) proved critical for maintaining engagement and reducing guilt-driven behaviors. Also, the ability to incorporate user-specific preferences, dietary choices, and methodologies (e.g. vegetarian diets, intermittent fasting, or glucose-stability principles) made the experience feel highly personalized and more relevant to individual needs.

During the booster, the goal is to design and test a Proof of Concept (POC) of the AI coach, focusing on real-time interaction and core use cases such as cravings, meal decisions, and recovery moments. The POC will be tested with clearly defined user groups, including individuals in life transitions (e.g. (during/after) pregnancy, going through an important weight loss program, etc..), professionals under stress, and health-conscious users who struggle with consistency. This includes developing a simple, functional prototype, testing it in real-life situations, and defining the core features for an MVP based on observed behavior and feedback. The key assumptions to validate could include: 1. Users will actively engage with the AI in the moment of decision (not only retrospectively or passively) 2. Users are willing to share real, imperfect eating behavior with an AI in a non-judgmental context 3. Real-time, adaptive guidance leads to different choices than users would have made alone 4. Small, suggested adjustments (rather than strict rules) are perceived as actionable and worth following. The main objective is to understand whether this approach becomes a natural part of users’ daily routines and meaningfully influences their behavior over time

The most important learning goal is to validate the key assumptions outlined in the previous phase, particularly whether users engage with the AI in real-life decision moments and whether this interaction does eventually influences their behavior in a consistent and sustainable manner. For instance, if users do not actively use the tool during key moments (e.g. cravings or meal decisions), or if the interaction does not influence their choices, it would indicate that the real-time companion model might not be sufficiently valuable or intuitive. In that case, the approach would need to be reconsidered. User feedback would be closely analyzed to identify whether there are overlooked opportunities, alternative user segments, or specific needs that require a different interaction model or value proposition. To ensure reliable insights, user interviews will be designed with open-ended questions, allowing participants to express genuine experiences and feedback rather than being guided toward expected answers.

Initial testing will be conducted with a small group of users within my personal and professional network, including individuals currently experiencing life transitions (e.g. pregnancy, new parents) and professionals under high stress. I am currently identifying potential partners in the health and wellness space, such as nutritionists, pregnancy support groups, and community health initiatives, to enable more structured testing. It would be valuable to receive support in connecting with relevant organizations that can provide access to diverse user groups and real-world testing environments.

The main goal is to validate and refine the concept in a structured, real-world setting.  In particular, access to relevant partners in the health and wellness space—such as nutrition experts, pregnancy support groups, and community health initiatives—would be highly valuable to test the solution with diverse user groups. Expert input in behavioral science and nutrition would further strengthen the approach and ensure scientific grounding, although initial contacts have already been identified. As the project is currently developed independently, the opportunity to connect with potential co-founders through the program would also be valuable to strengthen the team and support further development. Finally, connections to investors or organizations interested in supporting or participating in the initiative as early partners would be beneficial for the next stage of growth.

The project is currently led by myself, with a background in Product Management, Interaction Design, and AI-driven product development (including a strong experience in vibecoding). I am responsible for business and product development, design, as well as the prototyping and testing phase At this stage, I am supported informally by advisors with complementary expertise, including a Product Manager and a professional with a background in health and sports. Their input contributes to refining the product direction and grounding the concept in real-world use cases. As the project evolves, there is potential for these collaborators to take on a more formal role, depending on the development and opportunities identified.

To further develop the idea, collaboration with accredited experts in nutrition and behavioral science would be essential to strengthen the scientific foundation and credibility of the approach. In particular, formal collaboration with an established institution or certified professionals would help validate the methodology, ensure alignment with current research, and support long-term positioning within the health and prevention space. Additionally, expertise in AI system design and personalization would support the development of a robust and scalable interaction model. At the same time, I have the necessary competencies to develop a functional prototype with sufficient technical and AI performance to test the core concept (POC) and assess its potential.