By Michael J. Critelli | MakeUsWell Newsletter,
When a company brings a new product to market, it is natural to talk first about the technology, the features, or the problem it is trying to solve. All of that matters.
Our more important starting point here is how we are taking care before we ask people to use the product.
The purpose of our food intelligence product
At MoveFlux, we are preparing to beta test our AI-supported food intelligence product. It is designed to help people make smarter food decisions and gain better real-life insights at home, in grocery stores, in restaurants, while traveling, and in the everyday moments when food choices are made quickly and often with incomplete information.
Food is not just another consumer category. It touches health, family, culture, stress, sleep, work schedules, budgets, medications, allergies, and sometimes serious medical risks. A product that helps people think about food choices has to be useful, but it also has to be careful about what it claims, what it asks users to share, how it uses that information, and how clearly it explains its limits.
What we believe makes our product useful
Our product's value is not in precise calorie and nutrient calculations. We believe that the relationship between food and users is far more complicated, because it depends far more on individual situations and decisions.
Accordingly, we are looking at food intelligence at several levels:
The nutritional value of food as created
What happens to that nutritional value as food moves from origin through packaging, distribution, and preparation
How allergies and sensitivities affect the food experience
How food may interact with prescription drugs, over-the-counter medications, other foods, beverages, vitamins, and supplements
How social and cultural influences have both predictable and, at times, surprising, effects on nutrition decisions
Because of the interplay of these factors, what we need to know about our relationship with foods and what scientific studies reveal about them will keep changing.
The search for greater precision and certainty is a worthy one, but our product is based on the foundational belief that complete certainty is not achievable.
What makes sound nutritional habits is well established, but how each of us get there will vary.
A Thought Partner, Not a Physician
As we prepare to test and launch, we must be clear about our product’s role and its limitations.
Our product will help users have better conversations with physicians, nutritionists, pharmacists, and other trusted professionals about relationships among food, health, medications, allergies, and personal goals.
It is not designed to diagnose, treat, cure, prevent, or manage any disease or medical condition. Instead of replacing qualified professionals, our product serves as a thought partner to help users ask better questions, gain better insights, and make more informed decisions when working with the professionals they rely on.
Users still need to read labels, check ingredients, avoid known allergens, follow professional advice, and make final decisions for themselves.
Building on Personal Context Safely
That is why we are building with safety and responsibility in mind.
Our product becomes more useful when a user chooses to share personal context: preferences, allergies, sensitivities, goals, habits, meal timing, stress patterns, sleep, and travel. But if that information is not provided, the product may not account for it. And even when it is provided, no AI system can guarantee it will identify every risk or opportunity.
Privacy protection is a given
The more data a user shares, the more carefully we must explain what may be collected, how it may be used, and how it may help us improve the product. We will collect email addresses, but we assign each user a unique identifier so that all other personal data is anonymized and separated at an additional remove from direct human access.
We are minimizing AI-induced mistakes
We have designed a process that assigns tasks to multiple AI agents to reduce the risk of mistakes or misleading answers through fact-checking and research evaluation.
Because AI can rely on research that is uneven in quality, we have built a feature that rates the research our product draws on using rigorous evaluation criteria.
But the product might still produce inaccurate information. It may fail to account for context. And while we are training the output language to reflect appropriate levels of certainty, the output may seem authoritative even though incomplete.
The product’s features may change, especially based on beta user input. Some content may not fit a particular person's circumstances or may not be communicated as clearly as needed. These possibilities are not failures of the beta process; they are the reasons we are running it, providing the exact data points we need to learn, improve, and earn your trust.
We are focused on building trust
We view the beta process as a critical phase of product development, not a mere formality. We are reviewing this process with outside counsel not to generate excessive legal jargon, but to ensure that what we disclose is thoughtful, accurate, and fair. We want our disclosures to users to be truly informative and even educational.
During the beta, we may draw on session data to evaluate the product and improve how it works. That data is anonymized and not associated with any personal information.
We are excited about what food intelligence can become. And we believe how we introduce it matters.
The beta process is not just a step toward commercialization. It is a way to learn, listen, improve, and earn trust.
If we do this well, we will bring our product to market responsibly.