E-commerce
July 1, 2026
"Is the color true to the photos?" "Is it more matte or satin?" "What's the difference between midnight blue and black?" Three pre-purchase questions where the customer hesitates without a grounded answer, then opens a post-reception ticket or abandons the cart.
An e-commerce color material rendering AI chatbot does not replace COLOR-FLOW agents (#449). It reads COLOR-MAP, cites finish_note and lighting_disclaimer, compares similar colors, sends real_photo_url and routes checked-stock color exchange, then hands off post-reception to #449.
This guide #450 covers intents bot_color_*, flow COLOR-BOT and KPI color_bot. Distinct from color after-sales service (#449) and custom validation bot (#418): here, AI usecase to help choose: pre-purchase color, material and visual rendering.
Summary
Why does a color bot reduce returns related to shade matching?
An avoidable color ticket is born when the customer orders without a finish_note, lighting disclaimer, or color comparison before purchase.
Five botifiable color ticket triggers
Pre-fidelity: accurate color photos?
Finish ask: matte, satin, glossy, texture
Compare variants: navy vs midnight vs black
Material ask: composition, feel, care
Lighting context: natural light appearance
Digital Applied estimates that approximately 45% of fashion and home decor returns come from size, fit, or color/description discrepancy, often correctable upfront (Digital Applied, returns playbook 2026). Kn8 points out that a storefront agent verifies the live catalog before recommendation, unlike a static chatbot (Kn8, e-commerce agents 2026). Shopify allows images and swatches per color variant_id (Shopify, variants 2026).
Angle #450 vs neighboring content
After-sales color #449: COLOR-FLOW return exchange execute. The #450 = tier 1 pre-purchase fidelity compare bot.
Perso bot #418: embroidery thread color validation. The #450 = catalog colors, not custom-made.
Wrong item #186: picking error. The #450 = appearance expectations not fulfillment.
PDP blocks: static sheet FAQ. The #450 = contextual variant conversation.
Return analysis: dashboard reasons. The #450 = upfront deflection of color tickets.
Example of DTC home decor & textiles
Top 15 pastel colors 380 PDP sessions/week, 34% returns color reason baseline #449. After COLOR-BOT: color_bot_resolution 86%, color_bot_ticket_delta -47%, color_bot_pre_deflect 31%, color_bot_csat 4.5/5.
Variant gate CB-2
Verify product handle + selected color variant_id before COLOR-BOT. Material question without color context → bot_color_material_cite product level first.
Jewelry finish density
Plated gold rose gold : highest bot_color_finish_cite volume. Pre-load COLOR-MAP macro photos CB-6.
Seasonal launch pre-scale
New color collection : preload T1 T3 and COLOR-MAP extended disclaimer before launch day bot traffic spike.
Deco textile lot variation
bot_color_pre_fidelity cites supplier_batch metafield if color_lot_variation flag in COLOR-MAP. Sets realistic expectation pre-buy.
How does COLOR-BOT differ from COLOR-FLOW #449 and bot #418?
Material color bot, color agents playbook and custom validation bot: three layers on shared COLOR-MAP #449.
Bot color → role matrix
#450 COLOR-BOT: pre fidelity, finish cite, compare variants, real photo, exchange deeplink pre-buy
#449 COLOR-FLOW: post-reception hue disappointment, return execute, defect QA
#418 VALID-FLOW: personalization engraving thread color confirm
#186 wrong: variant shipped ≠ ordered
Upstream router
« True color », « matte or glossy », « navy black difference », « cotton composition », « real photo » → COLOR-BOT. « Received darker disappointed return », « dye stain » → handoff #449 CB-7. « Embroidery thread color » → bot_color_route_418. Order mismatch color → bot_color_route_186.
COLOR-BOT Data
COLOR-MAP JSON #449 finish_note lighting_disclaimer real_photo_url, Shopify variant images swatch, material metafield product, alternate color variant inventory CB-4.
Promise #450
COLOR-BOT policy, 12 intents bot_color_*, flow CB-1 to CB-8, guardrails no exact color guarantee, KPI color_bot_*.
Pre-purchase not post-return
#450 primary job = convert informed buyer. #449 primary job = resolve disappointed receiver.
bot_color_route_186 gate
Order line variant_id ≠ client claim shipped color: immediate handoff wrong item not COLOR-PRE macro.
Which bot_color_* intents should be configured?
Twelve color material rendering bot intents cover tier 1 pre-purchase and routing.
Twelve bot_color intents
bot_color_swatch_lookup: product + color → variant_id CB-2
bot_color_pre_fidelity: loyalty photos grounded disclaimer
bot_color_finish_cite: finish_note matte satin glossy
bot_color_compare_variants: navy midnight black side by side
bot_color_lighting_disclaimer: studio vs natural standard copy
bot_color_real_photo: real_photo_url unretouched send
bot_color_material_cite: composition care from product metafield
bot_color_alternate_suggest: alternate colors in stock CB-4
bot_color_add_cart_deeplink: /cart/add variant_id color chosen
bot_color_return_policy_cite: return_tier summary pre-buy only
bot_color_handoff_449: post-order hue defect disappointment
bot_color_route_418: personalization thread color
Session tags
color_bot, color_bot_pre, color_bot_compare, color_bot_resolved, color_bot_handoff_449, color_bot_route418. Distinct color_agent, perso_valid.
Triggers T1-T5
T1: PDP swatch select + dwell 10 s "Color questions" chip. T2: keyword true color material glossy navy. T3: swatch hover hesitation proactive. T4: collection page color filter context. T5: cart line color property review pre-checkout.
Intent priority
bot_color_swatch_lookup before bot_color_compare_variants. Never compare without CB-2 variant context.
P1 P2 intent routing
P1: bot_color_pre_fidelity, bot_color_handoff_449. P2: bot_color_compare, bot_color_finish. P3: bot_color_material FAQ.
How to apply the COLOR-BOT flow in eight steps?
The COLOR-BOT flow guide helps choose grounded colors COLOR-MAP #449.
Eight steps CB-1 to CB-8
CB-1 Welcome: "I can help you with the color and rendering of this product."
CB-2 Lookup variant: product handle + color label → variant_id
CB-3 Classifier intent: bot_color_* section 3
CB-4 Read COLOR-MAP: finish_note, lighting_disclaimer, real_photo, return_tier
CB-5 Read catalog: variant images, material metafield, alternate colors inventory
CB-6 Respond: fidelity | finish | compare | material | real photo
CB-7 CTA: add cart deeplink | alternate color | /pages/color-accuracy | handoff
CB-8 Close/log: color_bot_action, variant_color, pre_deflect, converted
CB-6 bot_color_pre_fidelity
COLOR-PRE-01 template from #449: finish_note + lighting_disclaimer + real_photo_url link + bot_color_return_policy_cite one line.
CB-6 compare branch
bot_color_compare_variants: cite COLOR-COMPARE-01 navy midnight black with swatch URLs from CB-5 variant images.
CB-6 material branch
bot_color_material_cite: product metafield composition + finish_note. Do not invent fiber percentages.
CB-7 alternate suggest
Customer hesitates on color: bot_color_alternate_suggest in-stock siblings + bot_color_add_cart_deeplink if accepts.
CB-8 pre_deflect log
Log session COLOR-PRE sent → purchase 7 d → no color return 30 d. Dashboard color_bot_pre_deflect.
CB-7 /pages/color-accuracy fork
Deep FAQ link for clients wanting full policy read. Bot summarizes then CTA page not wall of text.
CB-6 lighting branch standalone
Customer asks about lighting only: bot_color_lighting_disclaimer full COLOR-LIGHT-01 copy from #449 plus link real_photo_url practical comparison tip.
Which COLOR-BOT policy should be documented?
The COLOR-BOT color material policy incorporates COLOR-SUP #449 with pre-purchase AI guardrails.
Eight COLOR-BOT rules
COLOR-MAP only: finish disclaimer return_tier from JSON #449
Catalog grounded CB-5: material images from metafield API not LLM
No exact color guarantee: lighting_disclaimer always CB-6 fidelity
No agent opinion color: cite COLOR-MAP not "in my opinion"
Compare max 3 variants: sibling colorways same product only
Alternate verify stock CB-5: qty > 0 before deeplink
Max 8 turns color flow: then add cart CTA or handoff #449
Post-order handoff #449: bot never execute return label refund
Corpus RAG sources
/pages/color-accuracy, COLOR-MAP #449, COLOR-PRE-01 COLOR-COMPARE-01 templates section 7 #449, product material metafields.
Distinct #418 personal
Catalog swatch = COLOR-BOT. Line item embroidery thread = bot_color_route_418 immediately.
Monthly bot color review
Audit: exact match guarantee language, invented material %, compare wrong product, post-order return promise.
COLOR-SUP prerequisite gate
No COLOR-BOT prod until #449 COLOR-SUP published and /pages/color-accuracy live.
What guardrails protect accuracy and substance?
The authorized vs prohibited bot color matrix protects customer expectations and conversion.
Authorized bot actions
Cite finish_note lighting_disclaimer from COLOR-MAP
Compare up to 3 sibling color variants
Send real_photo_url if exists CB-4
Cite material metafield composition care
Suggest alternate colors in stock
Return_tier summary pre-buy cite only
Add cart deeplink chosen variant
Handoff post-order #449 pre-filled
Prohibited bot actions
Guarantee exact screen match
Invent material composition percentages
Blame client monitor aggressively
Execute return refund post-order
Compare cross-product colors
Route catalog color to #418 perso
Anti-hallucination catalog
Template fidelity: « [Color] : [finish_note]. [lighting_disclaimer]. » from COLOR-MAP JSON. See hallucination prevention.
Empathy without admission
Pre-buy: inform, do not apologize for brand fault. Post-order disappointment → handoff #449 COLOR-EMPATHY-01 agents.
Leather grain disclaimer
Natural leather: bot_color_material_cite unique grain per hide from COLOR-MAP extension. No two identical.
How to connect COLOR-MAP #449 and Shopify variant data?
The color-material bot integration combines COLOR-MAP, variant images and product metafields.
COLOR-BOT read fields
product_id: handle from PDP context
color_variants[]: label, variant_id, image_url, inventory_qty
selected_variant_id: from CB-2 swatch match
finish_note: COLOR-MAP per variant or product
lighting_disclaimer: COLOR-MAP standard or extended pastel
real_photo_url: unretouched stock image or null
material_metafield: composition care instructions
Shopify product metafields custom.material_composition (Shopify, metafields 2026).
PDP context T1
Pass selected swatch + product handle CB-2 pre-fill. Highest color_bot_pre_deflect conversion rate.
Variant image gallery CB-5
bot_color_compare_variants uses variant-specific image URLs not parent default only.
Qstomy Integration
See Shopify integration + COLOR-MAP RAG sync #449.
Order context post-pay
If order_id detected + hue disappointment keywords → bot_color_handoff_449 skip pre macros.
Swatch hex metafield CB-4
Optional hex_swatch COLOR-MAP enriches bot_color_compare_variants visual cite alongside image URLs.
What PDP, swatch, and UX triggers for COLOR-BOT?
The UX COLOR-BOT deployment intercepts color hesitation before abandonment.
Five widget placements
PDP swatch T1: "Question about this color?" chip after select
Finish note expand T2: "Learn more about material" opens bot
Compare hover T3: "Navy or Black?" proactive on dual swatch dwell
Cart color review T4: recap finish disclaimer pre-pay
/pages/color-accuracy T5: FAQ link opens bot context
Proactive message CB-1
"[Water green]: [finish_note]. Slight variation possible depending on screen: [lighting_disclaimer one line]. Real photo or compare another color?"
Pastel collection launch
Pre-load T1 T3 on new pastel SKUs with extended lighting_disclaimer before launch support spike.
Mobile swatch UX
Swatch tap opens bot drawer not dead end. Critical mobile conversion color mode.
Video 360 link CB-7
If product video metafield exists: bot_color_pre_fidelity includes video URL natural + studio light.
A/B T1 proactive vs passive
50% PDP color chip vs none 4 weeks min 600 sessions per hero colorway. Measure color_bot_pre_deflect lift.
T4 cart recap mobile
Cart drawer one-line finish_note before pay reduces color_finish_surprise post-reception on mobile majority traffic.
Finish note inline PDP sync
Merch metafield finish_note visible under swatch complements T1 bot chip. Dual channel reduces color_finish_surprise tickets 20-30% in pilots.
Which color bot KPIs should be measured?
The color material bot KPIs link pre-purchase deflection, conversion, and upstream returns.
Eight key metrics
color_bot_resolution_rate : resolved without handoff / color_bot sessions
color_bot_ticket_delta : decrease in tickets #449 vs baseline
color_bot_pre_deflect : post COLOR-PRE purchases without color return within 30 days
color_bot_compare_engagement : compare sessions / swatch dwell Q3
color_bot_add_cart_rate : deeplink click / sessions resolved pre
color_bot_handoff_449_rate : post-order routed agents
color_bot_real_photo_click : real_photo_url opens / offers
color_bot_csat : color_bot resolved satisfaction tag
DTC fashion home decor benchmark
color_bot_resolution 84-90 %, color_bot_ticket_delta -42-52 %, color_bot_pre_deflect 26-34 %, color_bot_add_cart_rate 38-48 %, color_bot_csat > 4.3/5.
Monthly dashboard
Trigger Q1-Q5 breakdown, top colors by sessions, compare vs fidelity intent split, handoff 449 rate, SKU color return correlation.
Cross-referencing returns analysis
Measure color return rate delta on SKUs with COLOR-BOT Q1 active vs control. Align returns analysis color reason.
SKU pilot color return correlation
Track color return rate 30 days pre vs post COLOR-BOT on top 10 pastel SKUs. Target -25 % minimum pilot cohort.
Which anti-patterns should be avoided on a color and material bot?
Twelve color rendering bot anti-patterns to ban.
1. Exact match guarantee
Rule 3 lighting_disclaimer mandatory. Legal and return dispute risk.
2. Invent material composition
Rule 2 CB-5 metafield only. Hallucination fiber %.
3. Post-order return execute bot
Rule 8 handoff #449. Bot pre-buy only.
4. Blame monitor pre-buy
Inform disclaimer not lecture client screen.
5. Skip COLOR-MAP #449
Deploy bot before finish_note documented.
6. Compare cross-product
Rule 5 sibling variants same PDP only.
7. Route catalog to #418
Swatch ≠ thread embroidery. bot_color_route_418 disambiguation.
8. Agent opinion color
Rule 4 COLOR-MAP cite only.
9. Ignore real_photo_url null
Offer variant images CB-5 fallback not "no photo".
10. No alternate when hesitating
bot_color_alternate_suggest missed conversion opportunity.
11. Generic « voir fiche produit »
COLOR-PRE-01 grounded beats defer.
12. Confondre wrong item #186
Pre-buy fidelity ≠ order mismatch. Route bot_color_route_186 if order context.
13. Deploy before COLOR-MAP
Error: bot live without #449 finish_note published. Hallucination and policy drift guaranteed.
How does Qstomy automate color and material?
Qstomy on Shopify runs COLOR-BOT: CB-2 swatch lookup, CB-4 COLOR-MAP read, bot_color_pre_fidelity, bot_color_compare_variants, bot_color_material_cite, real_photo send, alternate color deeplink, handoff #449 post-reception pre-filled CB-4 fields.
Qstomy bot_color capabilities
bot_color_map_rag: finish disclaimer return_tier
bot_color_pre_template: COLOR-PRE-01 auto PDP
bot_color_compare_engine: 3 variant side by side
bot_color_material_read: metafield composition
bot_color_cart_deeplink: variant add in-chat
bot_color_handoff_449: post-order ticket pre-fill
Pipeline #449 → #450
#449 COLOR-SUP COLOR-MAP prerequisite. #450 bot tier 1 deflect 85 %+ color pre-purchase. Shared policy post-reception agents.
Quantified DTC Scenario
Pastel home textile, 89 color tickets/month baseline #449, color_fcr 59%.
After Qstomy T1 swatch chip COLOR-BOT: color_bot_resolution 87%, color_bot_ticket_delta -48%, color_bot_pre_deflect 32%, color_return_rate pilot SKU -29%, color_bot_csat 4.6/5.
Explore customer support and request a demo.
Continuity merch CF-8
CB-8 sessions flagged high compare intent feed conversations → PDP photo updates.
Weekly color_bot transcript audit
Scan guarantee exact match phrases count zero. Material % invent count zero. Handoff 449 rate stable post-order.
RAG corpus color bot
Feed COLOR-MAP + 40 anonymized CB-8 transcripts for intent bot_color tuning. Quarterly review with ops #449.
What is the checklist for deploying COLOR-BOT?
COLOR-BOT Checklist (12 steps)
Validate COLOR-SUP #449 + COLOR-MAP /pages/color-accuracy
Export COLOR-MAP JSON + product material metafields
Configure 12 bot_color_* intents section 3
Implement flows CB-1 to CB-8 + swatch gate CB-2
Deploy T1 PDP swatch chip + T3 compare proactive
Guardrails no exact guarantee + material metafield only
Route pre-buy vs post-order handoff #449 disambiguation
Route custom #418 vs catalog color bot_color_route_418
Tests staging 8 scenarios: fidelity, finish, compare, material, alternate deeplink, post-order handoff, custom route, wrong 186
Monthly color_bot_pre_deflect KPI dashboard
A/B T1 proactive 4 weeks hero colorway
RAG sync COLOR-MAP on policy change
In brief
#450 = pre-purchase color material bot, not customer service #449 nor custom #418
COLOR-BOT: fidelity finish compare material deeplink
COLOR-MAP #449: finish disclaimer real_photo grounded
No exact guarantee: lighting_disclaimer always
KPI color_bot_pre_deflect: target > 28 %
FAQ
Bot vs agents #449?
#449 post-receipt return exchange defect execute. #450 bot tier 1 pre-purchase fidelity compare.
Difference bot #418?
#418 personalization thread color. #450 catalog color swatch.
Accurate color guaranteed?
No. bot_color_lighting_disclaimer + finish_note COLOR-MAP, exact match never promised.
Customer disappointed post-receipt?
bot_color_handoff_449 pre-filled ticket agents COLOR-DISCLAIMER-01.
Which PDP triggers?
T1 swatch chip, T3 compare dwell, T4 cart color recap pre-pay.
Going further
This week: validate COLOR-MAP #449, configure intents bot_color_*, deploy T1 PDP swatch chip, test CB-6 compare navy/black grounded images, connect material metafields CB-5, launch color_bot_pre_deflect dashboard.
Share this #450 guide with merchandising and support: a bot that quotes finish_note before purchase is worth ten returns of "not the same color", an exact match promise is worth a hundred disputes and zero swatch trust.
Go-live test matrix
8 scenarios sign-off: pre fidelity pastel, compare navy black, material cite, real photo send, alternate deeplink, post-order handoff, custom route 418, order mismatch 186.
Document CB-8 logs color_bot_action per scenario before production traffic enable.
Enzo
July 1, 2026
