AI Helps Tackle Allergy Risks

Key Concepts

Necessity-driven startup: A business founded to solve a problem the founder personally experienced.
Celiac disease: An autoimmune disorder triggered by gluten, requiring strict dietary avoidance.
Food allergies and dietary requirements: A broad category encompassing various sensitivities and preferences beyond common ones like gluten-free.
Large Language Models (LLMs): AI models trained on vast amounts of text data, capable of understanding and processing complex information.
Ingredient-level breakdown: Analyzing food items to identify each individual component.
Allergen tagging: Assigning specific allergen categories to ingredients.
Personalized menu solution: Tailoring menu information to individual dietary needs.
Hallucinations (in AI): When an AI model generates incorrect or fabricated information.
Structured ingredient and supplier data: Organized and verified information about food components and their sources.
Dietician team: Professionals who provide expert oversight and quality assurance for dietary information.
Legal liability: The responsibility for damages or injuries caused by a product or service.
Cross-contamination: The unintentional transfer of allergens from one food item to another.
Two-sided market: A platform that connects two distinct groups of users (e.g., restaurants and consumers).
QR code integration: Using quick response codes to link physical menus to digital information.
Dietary profile: A user's personalized set of dietary restrictions and preferences.
Senate Bill 68 (California): Legislation mandating detailed allergen information on menus for major restaurant chains.
Non-food origins: Refers to major allergens like gluten, dairy, nuts, etc.

Founder's Personal Journey and Problem Identification

Dylan's entrepreneurial journey is rooted in personal necessity. He was diagnosed with celiac disease at the age of ten, which led to extensive personal experience navigating the challenges of dining out and ordering food online. The difficulty in obtaining accurate information about food ingredients and the frequent inaccuracies he encountered became a significant source of frustration. This frustration motivated him to seek a solution that could help restaurants provide better allergen information to consumers.

The Limitations of Current Allergy Information Systems

While Dylan acknowledges that many restaurants have improved their understanding of common dietary needs like gluten-free, vegan, and vegetarian options, he highlights that these cover only a fraction of the population. He states that "there's 173 million Americans who have some form of food allergy or dietary requirements," and the allergens extend far beyond the most common ones. The current system, which relies on servers' knowledge and word-of-mouth communication, is demonstrably insufficient. A key statistic cited is that "54% of all allergic reactions in restaurants occur after the staff have been notified," underscoring the failure of the existing approach.

Dylan's Solution: AI-Powered Personalized Menus

The core of Dylan's solution involves leveraging technology to provide accurate and personalized allergen information. The process begins with restaurants providing their menu, recipe, and product information. This data is then ingested by their system, which utilizes trained large language models (LLMs). These LLMs are designed to "break those down to the ingredient level, tag them with the correct allergen and dietary requirements."

The output is a "personalized menu solution whereby consumers can see exactly what they can and can't eat on the menu depending on their personal requirements." This means that when a consumer accesses the menu, it is filtered and presented based on their specific allergies and dietary needs.

Addressing Liability and Ensuring Accuracy

A significant concern raised is the potential for errors and liability, especially given the life-threatening nature of some food allergies. Dylan, with his background as a former corporate attorney, addresses this by emphasizing the robustness of their technology.

No Guesswork: He states, "our technology never guesses if there is, you know, it's based on structured ingredient and supplier data." This means the AI relies on verified information, not assumptions.
Handling Uncertainty: In scenarios where there is uncertainty in the data, the system flags it. "If there's ever a scenario where it is insurer, it will tie in the back end for us that it's there's uncertainty."
Human Oversight: A dedicated "dietician team will come in over the top and do Q A and manually intervene." This human element ensures accuracy and provides a crucial layer of quality control.
Learning System: The AI learns from these manual interventions, becoming "smarter and smarter over time."

From a legal perspective, Dylan argues that the absence of documented allergen information presents a "much higher risk." He contrasts this with their system, which aims to remove the burden from untrained staff who are currently the "line of protection for the restaurant between the consumer and a potentially life threatening, a life threatening incident."

The Two-Sided Market and Integration Process

Dylan clarifies that their model operates as a two-sided market, requiring partnerships with both restaurants and consumers.

Restaurant Partnership: They collaborate with "restaurant chains, food service operators." The initial step involves ingesting their "menu recipe, product technology from various tech stocks."
Consumer Access: The most basic integration involves placing a "QR code in venue on physical menus and menu boards and a digital link on their website."
Personalized Experience: When consumers scan the QR code or access the digital link, they are prompted to "create their dietary profile." This profile allows them to select from "over 150 different allergens and dietary requirements."
Instant Menu Display: The system then "instantly it will show them, here's exactly what you can eat, here's what you can eat, what a modifier and what that modifier is, and here's what you can't eat and why." This provides a "completely personalized" dining experience.

Impact of Senate Bill 68 in California

The conversation highlights the significant impact of recent legislation, specifically "Senate Bill 68 in the state of California." This bill, effective July 26th, mandates that "every restaurant chain and food service facility with 20 plus locations nationwide, where at least one of those is in California to label all of their physical and digital menus for the major non-food origins."

Restaurants have two options to comply:

Physically annotate every menu item.
Use a digital solution, such as a QR code linking to a digital origin menu.

Dylan notes that "their strong preference is to use a digital mechanism," which has led to increased engagement with his company. He confirms that this legislation has "obviously a major step change for restaurants" and has "gotten a lot more involved then than we certainly were a few months ago, which is fantastic." Despite this legislative push, they continue to work with "independents, chains, food service, facilities of all types."

Conclusion and Key Takeaways

Dylan's venture addresses a critical unmet need for individuals with food allergies and dietary restrictions. By leveraging AI and structured data, his company provides restaurants with a robust and scalable solution for offering personalized menu information. The system prioritizes accuracy through LLMs and human oversight, mitigating liability risks for businesses. The recent legislative changes, particularly Senate Bill 68 in California, are accelerating adoption, signaling a broader shift towards greater transparency in food allergen information. The core takeaway is the power of technology to transform a complex and potentially dangerous aspect of dining into a safe and personalized experience for millions of consumers.