E-commerce

Guide for choosing dietary supplements: caution, disclaimers, and customer support

Guide for choosing dietary supplements: caution, disclaimers, and customer support

June 28, 2026

"Does this magnesium cure my cramps?" "Can I take your gummies with my anticoagulant?" "Which supplement to lose 5 kg before summer?" On a DTC supplement store, these messages arrive every day. A well-intentioned agent responding by gut feeling transforms a commercial question into unauthorized medical advice.

The assistance in choosing food supplements must inform about the product as formulated and authorized, refer to a healthcare professional as soon as the question goes beyond this scope, and never paraphrase a claim outside the EFSA register. EFSA recalls that the person responsible for placing the product on the market guarantees the safety of the supplement (EFSA, food supplements).

This guide #147 covers customer questions, authorized claims, disclaimers, cautious bots, and escalations. Distinct from regulated products (#119) (cross-cutting framework) and cosmetics (#146): here operational supplement advice with strict limits.

Summary

Why is recommending a supplement online riskier than a standard product?

The customer buys a promise of well-being. They expect a personalized response regarding their health. Your customer support is neither a doctor nor a pharmacist.

DTC supplement specificities

  • Regulated claims: EU register, exact wording

  • Daily dose: label only, no adjustment by the agent

  • Interactions: medications, pregnancy, pathologies

  • Influencers: customer mentions TikTok, expects support validation

  • Medicine boundary: therapeutic claim = ANSM/DGCCRF risk

France: placing on the market

Any placing on the FR market requires online declaration to the DGCCRF via Téléicare before marketing (DGCCRF, Téléicare). Support does not compensate for a non-compliant label.

DTC Example

Vitamin brand, agent replies "yes for chronic fatigue" on a ticket. Customer screenshot → report. REG-SUPP macros + training: zero unauthorized claim incidents in the 8 months post-deployment.

How does it differ from the general regulated products guide?

Two complementary levels, not redundant.

Regulated Products (#119)

Regulated (#119) covers cosmetics, food, and supplements: CAN/CANNOT, REG-* macros, agent training, locked bot. This guide #147 delves deeper into the supplement vertical: assistance in choosing, cautious quiz, range comparison, pre-purchase tickets with a focus on well-being.

Cosmetics (#146)

Cosmetics (#146): topical use, INCI, skin routine. Supplements: ingestion, Regulation 1924/2006, Directive 2002/46/EC, statements "do not exceed the daily dose".

Promise #147

Answer "which one to choose" questions quickly without prescribing, compare two SKUs based on facts, route to doctor/pharmacist at the right time.

What questions do customers ask about dietary supplements?

Mapping supplement support questions prioritizes bot intents and macros.

Pre-purchase product choice

  • « Magnesium bisglycinate vs citrate: which one is for me? »

  • « Can I combine your 3 products? »

  • « Suitable for vegetarians / gluten-free / halal? »

  • « Can a 12-year-old child take it? »

  • « How long before seeing an effect? »

Usage and safety

  • « Dosage: 1 or 2 capsules? » (→ label only)

  • « With my treatment X? » (→ healthcare professional handoff)

  • « Pregnancy / breastfeeding OK? »

  • « Stomach ache after taking »

Gorgias Tags

`supplement_choice`, `supplement_interaction`, `supplement_adverse`, `supplement_claim_question`. See pre-purchase objections, misunderstood products.

What claims exactly can the support cite?

The EU allegations register is the only source of health claims to support. Paraphrasing is prohibited.

EFSA / EU Claim Rules

  • Exact wording: copy validated label/PDP

  • Conditions of use: dose, target population displayed

  • No therapeutic claims: cures, treats, prevents disease

  • No UGC: customer testimonial ≠ authorized claim

Examples of safe vs. prohibited formulation

Authorized if on PDP: "Contributes to the normal function of the immune system" (vitamin C, register conditions). Prohibited: "Boosts your defenses against cancer", "Makes you lose 5 kg", "Replaces your antidepressant".

Comparison of two products

Compare galenic form, dose per capsule, ingredients listed, respective authorized claims. No "A is more effective for you" without an individual medical basis.

Macro comparison

"[Product A] contains [X] mg of [nutrient] per dose. Authorized claim: [PDP text]. [Product B]: [Y] mg, claim: [PDP text]. To choose according to your health situation, consult a professional."

Which disclaimers and REG-SUPP macros should be deployed on a daily basis?

The supplement disclaimers guide the agent and bot on each sensitive conversation.

Mandatory notices to recall

  • Do not exceed the recommended daily dose

  • Keep out of reach of young children

  • Does not substitute for a varied and balanced diet

  • Our advisors are not healthcare professionals

Macro library (Gorgias)

REG-SUPP-001: standard pre-response disclaimer. REG-SUPP-002: drug interaction → doctor/pharmacist. REG-SUPP-003: pregnancy/breastfeeding → mandatory medical advice. REG-SUPP-004: adverse effect → stop + pro + feedback + quality. REG-SUPP-005: factual product comparison. See support templates.

Placement

Macro footer, bot home, supplement page hub, post-purchase usage email. Markets: legal-validated translation, not DeepL alone.

How to offer decision support without prescribing?

A cautious supplements quiz directs to a SKU without diagnostic or personalized promise.

Authorized questions (5-6 max)

  1. General objective: energy, sleep, immunity, sport (closed list)

  2. Preferred format: capsule, powder, gummy

  3. Exclusions: gluten-free, vegan, caffeine-free

  4. Health red flag: pregnancy, treatment, pathology → handoff, no recommendation

  5. Age: < 18 → children's products only if the PDP provides for it

Mapping logic

Objective "sleep" + melatonin-free if not sold → filter SKUs tagged `sleep_support` with visible PDP claims. Result: "According to your criteria, here is [Product]: [authorized claim]. Consult a professional if you are undergoing treatment."

What the quiz does not do

Personalized dosage, guaranteed effect duration, 4-product stack without interaction check. See product quiz, bundles.

How to manage medications, pregnancy, and chronic conditions?

These topics trigger an immediate handoff, with no bot or junior agent exceptions.

Drug interactions

Anticoagulants + vitamin K, thyroid + iodine, PPIs + magnesium: support never validates the combination. Macro REG-SUPP-002 + "Discuss this with your doctor or pharmacist using the INCI list from the product sheet."

Pregnancy and breastfeeding

Unless there is an explicit PDP mention regulations-approved for this population, response: prior medical advice is mandatory. No "our pregnant customers take...".

Pathologies

Diabetes, hypertension, kidney failure, thyroid disorder: refer to doctor/dietitian. Do not suggest stopping medical treatment.

Sport and weight loss

Unauthorized performance or weight loss claims = polite refusal + PDP claims only. Care Europe reminds: supplement vs. drug classification under ANSM control if formulation is borderline (Care Europe, supplements France 2026).

How do you configure the add-on chatbot on Shopify?

The supplement chatbot distributes approved text, not generative health advice.

Locked corpus

  • PDP: composition, dose, claims, warning pictograms

  • REG-SUPP macros validated by legal

  • Range comparison guide (factual, not relative efficacy)

  • Metafields: vegan, gluten_free, age_min

Intents and guardrails

Auto OK: `ingredient_list`, `daily_dose_label`, `claim_cite_pdp`, `format_compare`, `certification_check`. Handoff: `drug_interaction`, `pregnancy`, `adverse_reaction`, `diagnose_symptom`, `dosage_change`, `weight_loss_promise`. Blocklist: cures, heals, treats, prescribes, 100% effective. Confidence 95%+ on claims; otherwise "check product sheet" + agent.

Governance alignment

See bot governance (#142), hallucinations (#123), bot limits (#124).

How do I handle an adverse event reported by a customer?

An adverse event report requires empathy, stop usage, quality escalation, and no minimization.

Support Procedure (6 steps)

  1. Empathy: "I am sorry you are experiencing this"

  2. Advise immediate discontinuation of the product

  3. Recommend consulting a doctor if symptoms persist or are severe

  4. Collect: SKU, batch number, date taken, symptom description, photos if applicable

  5. Offer return according to policy + commercial gesture if policy provides for it

  6. Escalate to the quality manager within 24 hours for traceability

Nutrivigilance

Regulatory manager assesses adverse event report to Anses if severe. Support documents verbatim and batch number, does not comment on medical causality.

Macro REG-SUPP-004

Includes stop usage, doctor, return, and batch number requested. No "it's normal for the first few days" without official PDP text.

What errors in supporting documents trigger disputes and reports?

Five supplement anti-patterns to ban in training.

1. Validating a TikTok claim

"Yes, the influencer is right, it detoxes the liver" without any registered health claim.

2. Adjusting the dosage

"Take 3 capsules if you are tired" vs. label dosage.

3. Stacking products without caution

Recommending 4 cumulative supplements without mentioning the risk of nutrient overdose.

4. Minimizing side effects

"Continue for another week" regarding headaches or nausea.

5. Bot inventing an ingredient

"Lactose-free" hallucination when the product description page mentions milk. See marketing customer service alignment, quality of responses.

How does Qstomy secure the process of choosing supplements?

Qstomy combines product guidance and health safeguards for DTC supplement brands.

Supplement Features

  • Locked PDP Corpus: official claims and dosage

  • Prudent Quiz: red flags → automatic handoff

  • SKU Comparison: factual, cited claims

  • Disclaimer Injection: REG-SUPP on sensitive intents

  • Health Handoff: pregnancy, medication, symptoms

  • Audit Export: claim conversations for QA

Quantified DTC Scenario

Micronutrition brand, 520 conv/month, 35% usage/selection questions. Qstomy deployment: 5-question quiz, health blocklist, synced macros. After 6 months: L1 comparison tickets -41%, health handoff 100% compliant based on 50-thread audit, zero unauthorized claims by agents, quiz completers conversion +18%.

Explore AI support, Shopify, request a demo.

What are the operational playbooks for supplements support?

Playbook 1: REG-SUPP macros (2 h)

Write 5 macros section 5. Regulatory manager validation. Publish Gorgias + bot corpus.

Playbook 2: prudent choice quiz (3 h)

5 questions section 6, red flag handoff, objective mapping → tagged SKU. Result disclaimer. Test 10 paths including 3 health handoffs.

Playbook 3: 1 h agent training

CAN/CANNOT supplements, 10 prohibited scenarios, 90% quiz. Certification before solo tickets `supplement_*`.

Playbook 4: monthly audit 30 conversations

Score: exact PDP claim, disclaimer present, handoff if health. 1 macro or corpus fix/week if discrepancy.

Playbook 5: adverse effect procedure (45 min)

6-step document section 9. Macro REG-SUPP-004. Quality contact + batch sheet template.

Useful linking

Providing good advice on a supplement is not about guiding a prescription: it is about helping to read the product sheet and knowing when to hand over to the doctor.

Enzo

June 28, 2026

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

Subscribe to the newsletter and get a personalized e-book!

No-code solution, no technical knowledge required. AI trained on your e-shop and non-intrusive.

*Unsubscribe at any time. We do not send spam.

Subscribe to the newsletter and get a personalized e-book!

No-code solution, no technical knowledge required. AI trained on your e-shop and non-intrusive.

*Unsubscribe at any time. We do not send spam.