E-commerce

AI Chatbot for product reformulation: explain what is changing and for whom

AI Chatbot for product reformulation: explain what is changing and for whom

July 1, 2026

“The bot says nothing has changed even though the taste is different.” “The AI promises the old formula is in stock.” “Chatbot advises an allergy sufferer without citing INCI.” Three failures where a miscalibrated reformulation bot denies the change, invents ingredients, or promises a return outside of policy.

An e-commerce product reformulation AI chatbot does not replace REFORM agents (#703). It reads REFORM-MAP, cites ingredients_changed_list, who_benefits_copy, and handoffs exceptional allergen returns to humans.

This guide #704 covers intents bot_reform_*, flow REFORMbot, and KPI reform_bot. Bot pair of the REFORM playbook (#703). AI use case: explaining change in composition, taste, texture, and for whom, without medical promises or invented old formulas.

Summary

Why automate rewriting with a bot?

"Did you change the recipe?", "what does the new INCI contain?" and "I want the old formula" arrive right from the first post-reformulation use. A calibrated bot cites REFORM-MAP change reason, ingredients changed, who benefits and stock transition without denying the change or promising the unavailable old formula.

What the bot resolves tier 1 reformulation

  • What has changed: ingredients_changed_list CHANGE-CITE map

  • Taste texture perf: performance_change_copy explain map

  • For whom: who_benefits_copy cite map

  • Stock transition: transition_stock_policy old new formula map

  • INCI allergen info: allergen_change_flag cite NO-MEDICAL

Composition questions often spike within 7 days post-purchase if PDP does not show an INCI comparison (WisePIM, ingredient transparency 2026).

REFORMbot vs REFORM #703, PKGCHGbot #702, PRODUPD #671 and ingredients #166

Five contents, five distinct product evolution moments.

Quick Matrix

Pipeline: #704 bot explain change tier 1 → #703 agents return exception allergen escalate.

Which bot_reform_* intents should be configured?

Eight reformulation bot intents mapped to reform_* typologies #703.

Eight bot_reform intents

  • bot_reform_whats_changed: ingredients perf CHANGE-CITE map

  • bot_reform_taste_texture: performance_change_copy explain map

  • bot_reform_allergen_ingredient: allergen_change_flag ALLERGEN-CITE NO-MEDICAL

  • bot_reform_who_benefits: who_benefits_copy cite map

  • bot_reform_transition: transition_stock_policy old new mixed formula

  • bot_reform_old_formula: old_formula_avail_policy NO-OLD-FORMULA-PROMISE

  • bot_reform_return_policy: return_policy_reform explain handoff

  • bot_reform_handoff: return exception allergen reaction → #703

Tier 1 auto: whats_changed, taste_texture, who_benefits, transition if SKU REFORM-MAP + guardrails.

bot_reform_handoff return execute allergen reaction medical → agents #703 payload order_ref sku intent.

How do you consume REFORM-MAP #703?

The bot reads REFORM-MAP #703: reform_program_id, sku_affected, old_formula_summary, new_formula_summary, ingredients_changed_list, allergen_change_flag, change_reason_copy, performance_change_copy, who_benefits_copy, transition_stock_policy, old_formula_avail_policy, return_policy_reform, customer_communication_copy.

Lookup grounded

  • CHANGE-CITE-BOT: ingredients_changed_list cite map verbatim

  • ALLERGEN-CITE-BOT: allergen_change_flag INCI cite if concerned

  • NO-MEDICAL-ADVICE-BOT: allergen direct to pro no diagnosis

  • NO-OLD-FORMULA-PROMISE-BOT: old_formula_avail_policy cite map only

  • TRANSITION-CITE-BOT: transition_stock_policy old new formula map

  • NO-RETURN-EXECUTE-BOT: no refund return change of mind bot only

  • ORDER-SKU-VERIFY-BOT: order API sku match before formula claim

  • REFORM703-HANDOFF-BOT: return exception allergen → #703 agents RF-7

Alignment anti-hallucination (#123): INCI change reason who benefits = REFORM-MAP whitelist only.

REFORMBOT-SUP policy in six rules

Six rules for responsible bot reformulation.

  1. REFORM-MAP-GROUNDED-BOT: ingredients perf reason from map only

  2. CHANGE-CITE-BOT: ingredients_changed_list cite reform confirmed

  3. ALLERGEN-CITE-BOT: allergen_change_flag INCI cite NO-MEDICAL-ADVICE-BOT

  4. NO-OLD-FORMULA-PROMISE-BOT: old_formula_avail_policy cite map only

  5. NO-RETURN-EXECUTE-BOT: change of mind return agents #703 only

  6. PKGCHG702-REROUTE-BOT: packaging only, formula unchanged → #702 PKGCHGbot distinct reformulation

Flow REFORMbot RFB-1 to RFB-8

Eight-step product reformulation bot flow.

  1. RFB-1 Classify: bot_reform_* intent detect post-use chat widget PDP

  2. RFB-2 Collect: order_ref sku lot photos product received

  3. RFB-3 REFORM-MAP: ingredients perf allergen who benefits transition return old

  4. RFB-4 Order lookup: order SKU API match lot batch formula version if tracked

  5. RFB-5 Guardrail: CHANGE ALLERGEN NO-MEDICAL NO-OLD-FORMULA NO-RETURN HANDOFF

  6. RFB-6 Respond: TPL-REFORMbot grounded customer_communication_copy

  7. RFB-7 Handoff or close: return exception allergen reaction #703 payload

  8. RFB-8 Log: intent reform_program_id tag reform_bot change handoff Y/N

Example TPL-REFORMbot-CHANGE

"Order [order_ref] [sku map]: reformulation [reform_program_id map] confirmed. [change_reason_copy map.] Changes: [ingredients_changed_list map verbatim.] [performance_change_copy map.] CHANGE-CITE-BOT."

TPL-REFORMbot templates and touchpoints

Four essential templates.

TPL-REFORMbot-CHANGE

Program [reform_program_id] [sku]: [change_reason_copy map.] Modified INCI: [ingredients_changed_list map verbatim.] [performance_change_copy map.] CHANGE-CITE-BOT.

TPL-REFORMbot-WHO

Reformulation [reform_program_id]: [who_benefits_copy map.] New formula [new_formula_summary map.] Old: [old_formula_summary map.]

TPL-REFORMbot-ALLERGEN

INCI [sku map]: [ingredients_changed_list allergens map.] [allergen_change_flag Y : ALLERGEN-CITE-BOT.] Consult your doctor if allergic. NO-MEDICAL-ADVICE-BOT.

TPL-REFORMbot-HANDOFF

Request [return change opinion / old formula exception / allergic reaction] forwarded to the team. Summary [order_ref sku photos intent]. REFORM703-HANDOFF-BOT.

Touchpoints

  • PDP reformulation banner widget: bot_reform_whats_changed proactive

  • Email loyal customers footer chat: bot_reform_who_benefits entry

  • Chat keyword taste recipe: bot_reform_taste_texture trigger

  • Return portal formula reason: bot_reform_return_policy handoff

Edge cases and reroutes

Five cases outside tier 1 bot standard reformulation.

Bot explain change tier 1. Return exception old formula allergen reaction → agents #703 REFORM-RETURN-01.

Essential reform_bot KPIs

Five REFORMbot steering metrics.

  • reform_bot_change_deflect: whats_changed resolved without agent

  • reform_bot_change_cite_rate: CHANGE-CITE-BOT / bot_reform_whats_changed

  • reform_bot_allergen_cite_rate: ALLERGEN-CITE-BOT / bot_reform_allergen_ingredient

  • reform_bot_old_formula_promise_violations: old formula promised bot audit target 0

  • reform_bot_handoff_rate: return allergen / total reform bot tickets

Target: reform_bot_old_formula_promise_violations 0 and reform_bot_change_deflect greater than 55%.

REFORMbot anti-patterns

Five common bot reformulation errors.

  1. Denying formula change: CHANGE-CITE-BOT ingredients_changed_list map

  2. Promising old formula: NO-OLD-FORMULA-PROMISE old_formula_avail_policy map

  3. Allergen medical advice: NO-MEDICAL-ADVICE-BOT refer pro

  4. Bot invented INCI: REFORM-MAP-GROUNDED-BOT whitelist only

  5. Confusing PKGCHG #702: packaging only distinct reformulation composition

REFORMbot with Qstomy

Qstomy on Shopify: detect bot_reform intent, REFORM-MAP RAG grounded, change who benefits explain, allergen cite NO-MEDICAL, handoff #703 return exception tier 2.

Pipeline: #704 bot explain change tier 1 → #703 agents return exception escalate quality legal.

Explore AI support and request a demo.

Checklist, FAQ and going further

REFORMbot Checklist (8 steps)

  1. Sync REFORM-MAP #703: RAG bot PDP reformulation banner widget

  2. Policy REFORMBOT-SUP: 6 rules CHANGE ALLERGEN NO-OLD-FORMULA NO-RETURN

  3. 8 intents bot_reform_*: flow RFB-1 to RFB-8

  4. 4 templates TPL-REFORMbot-*: CHANGE WHO ALLERGEN HANDOFF

  5. Order SKU API sync: ORDER-SKU-VERIFY bot agents test

  6. PDP INCI comparative embed: bot_reform_whats_changed proactive

  7. Red team 10 prompts: deny change old formula promised invented INCI medical advice bot

  8. Dashboard KPI: reform_bot_* section 9 change_deflect old_formula_violations

FAQ

Difference #703?
#703 = agents return exception old formula escalate quality legal. #704 = bot tier 1 explain change handoff.

Difference PKGCHGbot #702?
#702 = packaging redesign formula unchanged. #704 = composition taste texture changed REFORM-MAP.

Does bot refund taste change?
No. NO-RETURN-EXECUTE-BOT TPL-REFORMbot-HANDOFF #703 return_policy_reform.

Does bot advise on new allergens?
TPL-REFORMbot-ALLERGEN ALLERGEN-CITE-BOT NO-MEDICAL-ADVICE-BOT direct to professional.

Go further

This week: index REFORM-MAP RAG PDP banner INCI comparative, sync order SKU API bot, red team old formula promised bot violations.

Enzo

July 1, 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.