top of page

AI Methodology Guidelines

AI Plan Generation

Our AI system generates personalized nutrition and exercise plans using advanced language models with structured prompts. The AI incorporates medical guidelines, user health data, and safety protocols to create evidence-based wellness recommendations.

​

AI Methodology: Our AI system uses structured prompts trained on established medical guidelines and evidence-based research. The system incorporates user-specific health data and safety protocols to generate personalized wellness guidance while maintaining medical accuracy.

​

How it works (high-level)

1. Input: user profile + calculated goals (caloric budget, macronutrient targets, step/exercise targets, health conditions).

2. Prompting: a structured prompt (internal) instructs Gemini to produce a meal plan or exercise plan aligned with the provided targets.

3. Post-processing: plans are validated against safety limits, calorie budgets, and adjusted for specific health conditions.

4. Attribution & Citations: each plan includes links to the methodology page and source guidelines (see Sources section below).

 

A. AI Features and Safety Disclaimer

  • Certain features of the Platform utilise artificial intelligence (“AI”) technologies to generate responses, suggestions, and insights (“AI Outputs”). By using these features, you acknowledge and agree that AI Outputs are probabilistic in nature, generated based on patterns in data, and are not deterministic. Consequently, AI Outputs may not always be accurate, complete, reliable, unbiased, or current, and must not be treated as a substitute for professional advice, judgment, medical advice, diagnostic advice, treatment, or expertise from qualified professionals.

​

  • While Lillia seeks to enhance the performance of its AI systems, we make no representation or warranty that AI Outputs will be accurate, appropriate, reliable, or error-free. Users are expected to exercise discretion and independent judgment when relying on or acting upon AI Outputs, and must not base important decisions, particularly those related to physical or mental health emergencies, solely on such outputs. In such situations, users must seek immediate assistance from qualified professionals or emergency services.

​

  • You expressly acknowledge the inherent limitations of AI, including the possibility that AI Outputs may be inaccurate, incomplete, outdated, or otherwise unsuitable for your intended use. All AI Outputs are provided strictly on an “as-is” basis, without any warranties or representations of any kind, whether express or implied, regarding their accuracy, reliability, completeness, or fitness for purpose. You bear sole responsibility for any decisions made or actions taken in reliance upon AI Outputs. The use of such features requires the exercise of critical judgment and user discretion, and you agree to independently verify any information upon which you intend to rely with a qualified professional.

​

  • Certain AI features may also be powered by third-party providers, in which case the terms and policies of such providers may additionally apply. Lillia disclaims any responsibility or liability for the independent practices, operations, or content of such third-party providers. You agree and consent that any information provided may be processed by AI for further AI Output. 

​

  • To the fullest extent permitted by applicable law, Lillia, its affiliates, officers, directors, employees, and agents shall not be liable for any claims, losses, damages, or harm, whether direct, indirect, incidental, consequential, special, punitive, or otherwise, arising out of or in connection with your use of, or reliance on, any AI features or AI Outputs. This limitation of liability expressly includes, without limitation, any damages resulting from misuse of AI, misplaced reliance on its perceived capabilities, or any failure by you to exercise reasonable and prudent judgment.

​

B. Use of Third-Party AI Services​

  • We use Google Cloud Vertex AI to securely process user messages and relevant data in order to provide AI-powered insights. Google Cloud acts solely as a data processor and does not use this data to train its models.

 

  • What Data Is Shared

With your explicit consent, we may transmit:

  • Messages and prompts submitted within the App

  • Health and wellness information you choose to enter (e.g., symptoms, conditions, lifestyle, goals)

  • Context needed to generate relevant responses (e.g., prior messages in the session). Only the minimum required data is shared.

 

  • Who the Data Is Shared With

Data is shared solely with Google Cloud Vertex AI to generate responses. These providers process your data according to their privacy policies and only for the purposes of AI functionality.

 

  • User Permission

We transmit your data only after you provide explicit consent. If you do not consent, you may opt out and will not have access to AI-powered features. You can withdraw consent at any time.

​

We do not sell, monetize, or share users' personal data with third parties for advertising, profiling, or any other purposes. We use Google Cloud Vertex AI solely as a data processor to provide AI-powered features such as insights, forecasting, and conversational analysis. Google Cloud processes user messages and relevant data (such as transactions and balances) strictly on behalf of us under Google Cloud’s enterprise data processing terms. Google Cloud does not use our customer data to train foundation models or for model improvement. All AI processing occurs only after the user provides explicit in-app consent by checking the box for Privacy Policy and AI Methodology & Guidelines on the dedicated consent screen. We retain full control over user data and determine the purposes and means of processing. Google Cloud acts solely as an infrastructure provider and processor.

​

AI Limitations & Boundaries

Our AI system has clear limitations and does NOT provide:

1. Medical diagnosis: AI cannot diagnose medical conditions

2. Medication advice: AI does not recommend medications or dosages

3. Medical treatment: AI does not provide medical treatment recommendations

4. Emergency care: AI is not designed for emergency medical situations
​

Our AI approach:

1. AI provides general wellness and lifestyle guidance only

2. AI incorporates physician-recommended care plans when provided by users

3. AI generates lifestyle assessments based on user data and established guidelines

4. AI recommendations are designed to complement, not replace, professional medical care
 

Safety & Medical Disclaimer

Important: Our AI-generated plans are intended for general wellness and lifestyle guidance only. They are not medical diagnoses, treatments, or personalized medical advice. Always consult qualified healthcare professionals before making significant health changes.

Our AI system includes safety protocols that detect potential health concerns and recommend professional medical consultation when appropriate.
 

AI Training Sources & Citations

Our AI system is trained on and references established medical formulas and guidelines:


Core Medical Formulas

1. BMR Calculation: Harris-Benedict Equation - Standard metabolic rate calculation

2. BMI Calculation: CDC BMI Calculator and Guidelines
 

Health Monitoring Guidelines

1. Blood Pressure: American Heart Association - Blood Pressure Monitoring

2. Diabetes Management: American Diabetes Association - Treatment and Care

3. General Health Guidelines: CDC Blood Pressure Guidelines
 

AI Transparency: These sources and citations are easily accessible within the app through the "AI Guidance" or "Sources" section in settings or help menu.

bottom of page