E-commerce
July 1, 2026
"Is this product description written by an AI?" "Is the packshot image artificially generated?" "Your FAQ contains an error, who checks the automatic content?" Three tickets where the AI store content transparency lacks a clear support response.
The support for AI-generated ecommerce content questions covers product descriptions, images, editorial texts, and reporting inaccuracies. Distinct from chatbot governance (#142) and bot hallucinations (#123): here, the customer queries the content displayed on the store, not just the chat widget.
This guide #859 deploys the AIGENCON-SUP policy, flows AC-1 to AC-8, and the AIGENCON-MAP matrix. Customer service pair of the future transparency bot (#860).
Summary
Why does AI-generated shop content generate tickets?
DTC brands use AI to write product descriptions, SEO variations, lifestyle visuals, and FAQs. The customer notices a generic tone, a factual error, or an image that is too polished. The agent does not know which content is AI-generated, downplays the issue, or redirects to the chatbot without a disclosure policy.
Five typical frictions in store AI content
AI description?: asks if the product text is automated
Generated image?: doubts about an AI photo or visual
Factual error: incorrect composition, size, or compatibility
Missing label: wants visible "AI-assisted content" badge
General distrust: no longer trusts website information
The EU AI Act imposes clear transparency when the customer interacts with an AI system (EU AI Act, Article 50). Support translates this obligation into ticket responses.
DTC Example
DTC Cosmetics, 9 aigencon_ tickets/month. After AIGENCON-MAP: aigencon_trust_resolution_rate 89%, error reports routed to merch 92%.
AIGENCON #859 vs Governance #142, Hallucinations #123 and bot #860
Six pieces of content, six distinct AI transparency angles.
Quick matrix
#859 AIGENCON: questions about AI content, shop descriptions, images, FAQ
Governance #142: internal rules, validation, bot supervision
Hallucinations #123: technical prevention of invented bot responses
Bot #860: should we say when a chat response is AI
Privacy #154: transversal data usage distinct from displayed content
Product questions: factual product question distinct from content origin
#859 = is this text or visual generated by AI? #860 = is your chat response AI?
Promise #859
Policy AIGENCON-SUP, tree AIGENCON-GATE, 8 macros, agent-ready AI content registry, KPI aigencon_trust_resolution_rate.
Which typologies does the aigencon_* classifier use?
Action-oriented classifier: disclosure ≠ merch correction ≠ bot handoff #860.
Eight AIGENCON-MAP typologies
aigencon_description_ask: AI-generated product description?
aigencon_image_ask: Artificially generated image or visual?
aigencon_accuracy_error: Factual error in AI content
aigencon_label_request: Wants visible AI transparency label
aigencon_human_review: Requests human review of content
aigencon_trust_general: General distrust of automated content
aigencon_editorial_faq: AI-assisted FAQ or blog page
aigencon_chat_confusion: Confusion between website content and chatbot response
AIGENCON-SUP Policy: Agent Rules and Escalation
The AIGENCON-SUP policy establishes honest transparency without over-promising zero AI or hiding assistance.
Six AIGENCON-SUP Rules
Up-to-date AI content registry: AIGENCON-DISCLOSE macro before improvisation
Disclose what is assisted: transparency policy cites internal map
Factual error → merch: AIGENCON-CORRECTION P2 product ticket
Human verifies: AIGENCON-HUMAN-REVIEW editorial process explained
Don't confuse chat and sheet: aigencon_chat_confusion → handoff #860
Don't deny AI if assisted: proportionate transparency registry
Content matrix (agent)
Product description: human-written, AI-assisted, or documented mixed
Images: real photo, retouching, or AI-generated according to registry
Blog FAQ: editorial content with or without assistance
Chat response: handoff transparency bot #860
Flow AC-1 to AC-8: standard resolution
Eight sequential steps, P3 SLA transparency < 24 h, escalate merch if accuracy_error.
Flow AC-1 to AC-8
AC-1 Triage: read question, tag aigencon_*, help center or chat?
AC-2 Lookup: AI content registry, affected SKU, content source
AC-3 Educate: AIGENCON-POLICY site transparency link
AC-4 Classify: aigencon_* via AIGENCON-MAP
AC-5 Execute: DISCLOSE, CORRECTION, HUMAN-REVIEW, handoff #860
AC-6 Confirm: macro AIGENCON-DONE exact scope
AC-7 Test: customer understands content origin and next step
AC-8 Close: KPI aigencon_trust_resolution_rate
Eight ready-to-paste AIGENCON-* macros
Factual macros aligned with internal register, no denial nor oversold "100% human".
AIGENCON-* Library
AIGENCON-DISCLOSE: "This content is {{status}}: written by our team, AI-assisted, or mixed. Details: {{policy_link}}."
AIGENCON-POLICY: "Our AI transparency policy explains which content is assisted. Link: {{url}}."
AIGENCON-ACCURACY: "Thank you for the report. We are forwarding it to the product team for verification within {{sla}}."
AIGENCON-HUMAN-REVIEW: "Sensitive sheets undergo human proofreading before publication. Priority reporting."
AIGENCON-LABEL: "We display a label on AI-assisted content when applicable."
AIGENCON-CORRECTION: "Error reported on {{sku}}. Product ticket {{ticket_id}} opened."
AIGENCON-LIMITS: "AI helps with writing. Regulated claims are validated by humans."
AIGENCON-DONE: "Summary: {{action}}. Contact us if other content raises questions."
AIGENCON-GATE tree and registry of AI agent-ready contents
Decision tree before denying AI or ignoring factual error.
AIGENCON-GATE
Question description or image? → DISCLOSE registry + POLICY
Factual error reported? → ACCURACY + CORRECTION merch P2
Label request? → LABEL + POLICY
Chat response confusion? → handoff #860 not DISCLOSE sheet alone
General mistrust? → POLICY + HUMAN-REVIEW + LIMITS
Minimum internal registry
Helpdesk table: content type, SKU or page, human/AI-assisted/generated status, review date, merch owner. Monthly update. Train agents: transparency ≠ admitting every error as "normal AI."
KPI, QA and handoff to bot #860
Measuring AIGENCON detects AI denial and under-routing of merchant corrections.
Four AIGENCON KPIs
aigencon_trust_resolution_rate: customer understands content origin / total
aigencon_correction_routed_rate: % accuracy_error with merchant ticket
aigencon_wrong_silo_rate: % routed to #860 or #123 incorrectly alone
aigencon_repeat_7d: same transparency question within 7 days
Bot #860 Handoff
Export AIGENCON-MAP to intents bot_aigencon_disclose, bot_aigencon_accuracy. Guardrail AIGENCON-SITE-VS-CHAT-BOT: distinguish store content from widget response.
Edge cases: regulated claims, UGC, multilingual content
Three cases out of standard flow.
Health or compliance claims
LIMITS + priority HUMAN-REVIEW. Legal escalation if sensitive claim is disputed.
Customer reviews or UGC
Not AI brand content. Explain distinction between customer reviews vs product sheet DISCLOSE.
Machine translation
Specify AI-assisted translation distinct from writing. DISCLOSE by locale if the registry documents it.
Agent training: 20 minutes AIGENCON
Module: content registry, honest transparency, merch CORRECTION, site ≠ chat #860.
Exercises
Ticket A: AI description? → DISCLOSE registry, not denial
Ticket B: composition error → merch CORRECTION, not "it's the AI"
Ticket C: incorrect chat response → handoff #860, not AIGENCON sheet
How Qstomy structures AIGENCON in your stack
Qstomy route aigencon_*, injects the AI content register and handoff #860 if chat confusion occurs.
Three building blocks
Routing: intent ai_content_transparency vs bot_hallucination vs product_question
Knowledge: AI content register synchronized with AIGENCON-* macros
Bot #860: tier 1 chat response transparency
Scenario: decor brand, 8 tickets/month on AI transparency. Bot #860 handles chat, agents process product sheet errors. aigencon_trust_resolution_rate goes from 72% to 90% in 5 weeks.
FAQ and AIGENCON deployment checklist
FAQ
Is the entire site AI-generated?
No. DISCLOSE by content type according to register. LIMITS on regulated claims.
Difference #860?
#859 = displayed shop content. #860 = AI-generated chat widget response.
Customer reports an error?
ACCURACY + CORRECTION merchant ticket. Do not minimize as an acceptable AI bug.
7-day Checklist
D1: AIGENCON-SUP + AIGENCON-MAP + AI content register
D2: 8 helpdesk macros
D3: public AI transparency policy page
D4: 20 min agent training
D5: aigencon_* tags + KPIs
D6: handoff test #860 vs DISCLOSE sheet
D7: bot brief #860 SITE-VS-CHAT-GATE
Linking
Enzo
July 1, 2026
