E-commerce
June 29, 2026
Questions about allergies and dietary restrictions are sensitive. A customer may ask if a product contains an allergen, if it is suitable for an intolerance, or if it is safe for a child. The response must be clear, but it must never become medical advice.
An AI chatbot can help retrieve ingredients, explain the notes present on the product sheet, and direct to a human when the question becomes too personal or risky.
This guide explains how to build a helpful chatbot on this topic: reliable information, cautious language, human handoff, and boundaries never to be crossed.
Summary
Why do allergen questions require so much caution?
A wrong answer about an allergen can have serious consequences. The customer is not just looking for product information: they want to know if the purchase is compatible with their health, their diet, or that of a loved one.
The chatbot can help if it sticks to factual ground: list of ingredients, declared allergens, possible traces, displayed certifications, and links to the product sheet. It must not state that a product is "risk-free" for a given person.
Good limit: the bot informs about the product. It does not decide if the product is medically suitable for a person.
Convert over 2,000 customers on average per month with Qstomy.
The world’s 1st Shopify AI dedicated to customer conversion
Empowering 200+ e-commerce merchants
What is the difference between product information and medical advice?
The product information describes what the brand knows and can prove: ingredients, allergens, labels, potential traces, manufacturing methods, or supplier manuals. Medical advice interprets this information for a specific individual.
The chatbot can answer: "This product contains nuts" or "The specifications indicate a possible trace of milk". It must not answer: "You can consume it despite your allergy".
This distinction must be visible in the wording. The bot must invite the customer to consult a healthcare professional in case of significant doubt.
What types of requests should be recognized?
The bot must distinguish questions about a specific allergen, a dietary restriction, a chosen diet, a certification, cross-contamination, or use for a vulnerable person.
Specific allergen: the customer asks if the product contains peanuts, milk, gluten, soy, or another ingredient. Dietary restriction: they are looking for vegan, halal, kosher, lactose-free, or gluten-free. Possible trace: they want to know if the factory manufactures other high-risk products. Sensitive case: they mention a severe allergy, a child, pregnancy, or a previous reaction.
Sensitive cases must be treated with greater caution and are often transferred to a human.
What data should be connected to the chatbot?
The chatbot must rely on validated sources: product sheets, ingredient lists, declared allergens, instructions, certifications, and available traceability information.
This data must be up to date. If a recipe changes, the chatbot must follow the new version. A food or cosmetic product cannot be handled with an approximate information base.
When the information does not exist or is not confirmed, the correct response is to say so clearly and forward the request.
Which response rules should be applied?
The bot must respond with facts and avoid any absolute promises. Sentences like "no risk", "totally safe for you", or "you can take it" are to be avoided.
A safer phrasing is: "According to the product sheet, this item contains [ingredient] and indicates [warning]. If you have a severe allergy or a medical doubt, we recommend seeking professional advice before purchasing."
Transparency is preferable to a reassuring but overly broad answer.
Which conversational flow should be followed?
The flow must secure the response without unnecessarily lengthening the conversation.
Identify the relevant product.
Understand the allergen or restriction requested.
Read the product sheet and official mentions.
Respond with the available information, without medical interpretation.
Transfer if information is missing or if the case is sensitive.
Which messages should be used?
For a declared allergen: "According to the product sheet, this item contains [allergen]. We advise you not to order it if this ingredient is problematic for you."
For a possible trace: "The sheet indicates a possible presence of traces of [allergen]. If your allergy is severe, it is preferable to seek medical advice or choose a product without this mention."
If the information is missing: "I do not have confirmed information on this point. I am forwarding your request to our team for verification before you order."
When to transfer to a human?
Transfer is necessary if the customer mentions a severe allergy, a past reaction, a child, a pregnancy, an illness, a treatment, or information missing from the product sheet.
The bot must transmit the product, the requested allergen, the notice visible on the sheet, and the customer's exact question. This prevents the agent from restarting the exchange.
Which KPIs should be monitored?
Track questions by allergen, forwarded requests, product sheets with missing information, and conversations where the customer drops off due to a lack of a clear answer.
If the same question comes up often, the product sheet needs to be improved. The chatbot should not permanently compensate for insufficient product information.
Which mistakes should be avoided?
Avoid absolute promises, diagnoses, personalized consumption advice, and invented answers when information is lacking.
Also avoid hiding uncertainty. Saying "I cannot confirm this point" is more reliable than giving an approximate answer on a health topic.
How can Qstomy help?
Qstomy can retrieve product information, respond with validated mentions, and transfer sensitive cases to a human.
The bot can also flag sheets where allergens or dietary restrictions are not clear enough, in order to improve product content.
Explore AI support or request a demo.
Enzo
June 29, 2026
