Start your supplement chatbot evaluation with questions a shopper could ask before buying. The fifteen examples below cover products and store information. Each includes the source an answer needs. Adapt the wording to your range and use the list as a test sheet. These are illustrative prompts, rather than transcripts or a claim that every answer is already present in your catalog.
Questions 1–3: what is in the product?
These questions need the current product’s ingredient and label information. A broad description of an ingredient cannot establish the contents of a particular bottle. Keep the exact SKU and pack version with the source you use to verify the reply.
The answer should identify the relevant statement and explain unfamiliar wording when helpful. Where the source is incomplete, record what needs checking. This gives the catalog owner a specific correction to make.
- “What ingredients are in this product?” Check the full ingredient statement, including the information supplied about the capsule shell.
- “Does this Omega-3 list EPA and DHA separately?” Check the named product’s label and preserve the units and serving basis.
- “Does the brand describe this product as vegan?” Look for an explicit statement for that product; an omitted word does not establish suitability.
Questions 4–6: which format and pack?
A format preference often gives the assistant a useful starting point. Someone asking for a powder has supplied a concrete shopping constraint. Compare the available listings and explain which options meet it.
Product formats and pack sizes may be separate products or variants, depending on how the merchant has arranged the catalog. Check the selected listing and its options before presenting a choice.
- “Do you have this in softgels?” Verify the format on the exact item and show the matching product or option.
- “How many capsules are in the bottle?” Use the pack quantity and distinguish it from the number in a labelled serving.
- “Is there a smaller pack?” Check the available variants or separate listings and show the actual pack names.
Questions 7–9: help me narrow the options
These prompts ask the agent to use shopper preferences. The recommendation should explain its selection in ordinary buying terms: the requested format, a stated budget or a documented product attribute.
Keep a record of the constraint and the result. If no item meets all the requested conditions, the assistant should describe that clearly and ask whether the shopper wants to change a preference.
- “Show me a powder within my budget.” Apply the customer’s stated amount to the prices available for the relevant store context.
- “What is the difference between these two bottles?” Compare the named items using the same fields, such as format and pack quantity.
- “Can you show an option without this listed ingredient?” Verify the complete available product information and flag any missing evidence.
Questions 10–12: what am I adding to the cart?
At this stage the customer needs a precise product choice. Make the selected pack, quantity and displayed price easy to check. A product-family description may be too broad for the actual purchase.
Test the cart interaction separately from the text. A reply that says an item was added should agree with the resulting cart. The shopper remains in control of selecting an option and taking the purchase action.
- “What does this pack cost?” Check the displayed price and currency for the selected option.
- “Is the larger size available?” Check that specific option rather than assuming every pack shares the same availability.
- “Can I add the smaller bottle?” Present the correct choice and a supported cart action, then verify the line item in testing.
Questions 13–15: what happens around the purchase?
Store information needs its own approved sources. A product listing may contain ingredients while a separate policy explains delivery or returns. Give the assistant the pages that contain the answers your team wants shoppers to receive.
Obsedia’s Supplements AI Agent brings product knowledge and supported store-policy answers into the conversation. Use these final questions to check how the agent applies your store information. Access to a customer’s private order is a separate capability that this list does not assume.
- “Do you ship to my country?” Check the merchant’s published delivery policy and any limits relevant to the destination.
- “Where can I read the delivery terms?” Provide the approved policy page and a short explanation of the relevant section.
- “What is your return policy for this item?” Use the published terms and identify any product-specific conditions that need checking.
Turn the prompts into a working answer set
Choose the questions that recur in your store and add the exact source your team would use for each answer. Keep the verified product or policy page beside the prompt. This makes the list useful for preparing knowledge and for checking an agent after a catalog update.
Start with a handful of well-documented questions, then add a missing-detail case. Use the Obsedia supplement demonstration to inspect the conversation style and bring your answer set to a store-specific evaluation. Keep a note of the questions your current product pages still leave unanswered.
Use these fifteen prompts as a source checklist. Every answer should have an identifiable product or store page behind it, and every purchase action should lead to the intended item.
