E-commerce
July 1, 2026
"The bot told me my product is a counterfeit without safe proof." "AI promises an immediate refund." "Chatbot confirms authentic even though the serial is not in the registry." Three failures where a poorly calibrated authenticity-doubt bot creates legal liability or false security.
An AI authenticity-doubt chatbot does not replace FAKE agents (#567). It reads FAKE-MAP, cites verify_url auth_markers official_channels, clarifies bot limits, and hands off investigate refund to humans.
This guide #568 covers intents bot_fake_*, flow FAKEbot, and KPI fake_bot. Companion bot to the FAKE playbook (#567). Sensitive AI use case: establishing limit proofs, escalating without confirming fake, nor promising refund.
Summary
Why automate authenticity doubts using a bot?
"How do I know if it's real?", "where can I buy safely?", and "this packaging looks suspicious" are common pre- and post-purchase questions for premium buyers. A calibrated bot references verify_url, auth_markers, and official_channels from the FAKE-MAP without deciding on a fake status alone.
What the bot resolves at tier 1
Verification guide: verify_url process steps
Auth markers: hologram, packaging, serial map
Official channels: site, stores, authorized dealers
Serial lookup: registry status if serial is provided
Purchase source: direct vs third-party guidance
BCG points out that 62% of luxury customers fear poorly applied AI on trust issues (BCG, luxury experience 2026). The bot intercepts bot_fake_verify_how before a fake is confirmed or invented.
FAKEbot vs FAKE #567, AUTHCERTbot #566, NUMEDbot #564 and COA scope
Five pieces of content, five angles of trust and proof.
Quick Matrix
#568 FAKEbot: bot tier 1 doubts proofs limits handoff
#567 FAKE: agents investigate refund takedown legal
#566 AUTHCERTbot: bot COA verify duplicate distinct
#564 NUMEDbot: bot serial collector rarity
Bot COA scope: #566 reroute if authcert COA only intent
Pipeline: bot explain verify limits → #567 investigate refund execute. #566 if question COA document only.
Which bot_fake_* intents should be configured?
Eight bot authenticity doubt intents mapped to typologies #567.
Eight bot_fake intents
bot_fake_verify_how : verification process before purchasing
bot_fake_markers : auth_markers packaging serial
bot_fake_channels : official_channels where to buy
bot_fake_serial_verify : verify_url lookup serial
bot_fake_source_clarify : direct vs third_party scope
bot_fake_marketplace_info : reporting process cite map
bot_fake_limits : bot limits no refund no fake confirm
bot_fake_handoff : doubt refund legal → #567
Tier 1 auto : verify_how, markers, channels, serial_verify, source_clarify, marketplace_info if FAKE-MAP + guardrails.
bot_fake_handoff, refund claim direct doubt → agents #567 payload source order_ref.
How should FAKE-MAP #567 be consumed?
The bot reads FAKE-MAP #567: fake_sku_scope, verify_url, auth_markers, official_channels, investigate_steps, refund_if_confirmed, marketplace_takedown, legal_contact.
Lookup grounded
SKU scope lookup: fake_sku_scope PDP category
FAKE-MAP row: markers channels verify investigate
Registry API: verify_url serial status auth only
NO-FAKE-CONFIRM-BOT: bot never declares fake without human investigate bird
NO-REFUND-BOT: bot never promises refund
LIMITS-DISCLOSE-BOT: bot limits cited each doubt session
Alignment anti-hallucination (#123) and AI governance (#142): auth status = registry only.
FAKEBOT-SUP policy in six rules
Six bot rules when doubts about authenticity exist.
FAKE-MAP-GROUNDED: verify markers channels from map only
AUTH-VERIFY-BOT: serial status via verify_url registry only
NO-FAKE-CONFIRM-BOT: no declaration of counterfeiting by bot alone
NO-REFUND-BOT: no refund promise by bot
SOURCE-CLARIFY-BOT: direct vs third party before guidance
INVESTIGATE-HANDOFF: doubt direct refund → agents #567
Flow FAKEbot FKB-1 to FKB-8
Eight-step flow for bot doubts about authenticity.
FKB-1 Classify : bot_fake_* intent detect
FKB-2 Source collect : direct third_party marketplace report
FKB-3 FAKE-MAP : verify_url markers channels
FKB-4 Registry lookup : verify_url API if serial_verify
FKB-5 Guardrail : NO-FAKE NO-REFUND LIMITS-DISCLOSE
FKB-6 Respond : TPL-FAKEbot grounded
FKB-7 Handoff : doubt refund legal → #567
FKB-8 Log : intent source tag fake_bot
TPL-FAKEbot-LIMITS Example
“I can guide verification via [verify_url] and markers [auth_markers]. I cannot confirm counterfeits or issue refunds: agent investigation if direct purchase [order_ref]. Official channels: [official_channels].”
TPL-FAKEbot templates and touchpoints
Four essential templates.
TPL-FAKEbot-VERIFY
Verification: [verify_url]. Markers: [auth_markers]. Serial [serial]: [registry status if lookup]. Official channels: [official_channels].
TPL-FAKEbot-SOURCE
Source: [direct / third-party / report]. [Direct: handoff investigate #567.] [Third-party: guidance verify only, no CS brand refund.] [Report: marketplace_takedown process legal ops.]
TPL-FAKEbot-LIMITS
Bot limits: no fake confirmation, no refund. Official verification: [verify_url]. Agent if doubt about direct purchase or reported URL.
TPL-FAKEbot-HANDOFF
Request [doubt direct / refund / legal] transferred to agent. Source [source]. Order [order_ref]. SLA [handoff_sla].
Touchpoints
Page /authenticity: bot_fake_verify_how embed
PDP premium: bot_fake_markers chip
Post-purchase doubt: bot_fake_handoff entry
FAQ counterfeit: bot embed FAKE-MAP RAG
Edge cases and reroutes
Five cases outside tier 1 bot verify standard.
Direct refund required: NO-REFUND-BOT handoff #567 FAKE-REFUND-01
COA document only: reroute #566 bot_authcert intents
Serial collector: #564 NUMEDbot overlap
Legal threat: INVESTIGATE-HANDOFF #567 fake_legal_escalate
Second-hand third parties: verify guidance only SOURCE-CLARIFY
Bot never investigates nor refunds: doubt direct refund takedown → agents #567 ops legal.
Essential fake_bot KPIs
Five FAKEbot steering metrics.
fake_bot_verify_deflect: verify sessions without a human ticket
fake_bot_limits_disclosed_rate: doubt sessions with LIMITS cited
fake_bot_fake_confirmed_invented: fake declared bot, target 0
fake_bot_refund_promised: refund promised bot, target 0
fake_bot_handoff_rate: sessions → #567 agents
Monthly red team: 10 prompts "confirm it's fake", "refund me", "it's authentic promise it". Fake confirmed invented = NO-FAKE-CONFIRM-BOT violation.
FAKEbot anti-patterns
Six common mistakes.
Fake confirmed bot : NO-FAKE-CONFIRM-BOT
Refund promised : NO-REFUND-BOT
Auth without registry : AUTH-VERIFY-BOT
Limits not mentioned : LIMITS-DISCLOSE-BOT
Third-party refund guidance : SOURCE-CLARIFY-BOT
COA confusion : reroute #566 AUTHCERTbot
FAKEbot with Qstomy
Qstomy on Shopify: detect bot_fake intent, FAKE-MAP RAG, verify_url lookup, NO-FAKE NO-REFUND guardrails, LIMITS-DISCLOSE template, handoff #567 investigate payload.
Pipeline: bot tier 1 verify limits → #567 investigate refund takedown → #142 governance audit fake bot.
Explore AI support and request a demo.
Checklist, FAQ and going further
FAKEbot Checklist (8 steps)
Sync FAKE-MAP #567: RAG bot + verify_url registry API
Policy FAKEBOT-SUP: 6 NO-FAKE-CONFIRM-BOT rules
8 intents bot_fake_*: flow FKB-1 to FKB-8
4 templates TPL-FAKEbot-*: VERIFY SOURCE LIMITS HANDOFF
Page /authenticity embed: bot_fake_verify_how entry
LIMITS-DISCLOSE: mandatory template for doubtful sessions
Red team 10 prompts: confirmed fake promised refund invented auth
Dashboard KPI: fake_bot_* section 9
FAQ
Difference #567?
#567 = agents investigate refund takedown legal execute. #568 = bot tier 1 proof limits handoff.
Bot confirms counterfeit?
No. NO-FAKE-CONFIRM-BOT. Handoff #567 investigate.
Bot refunds?
No. NO-REFUND-BOT. Handoff #567.
Difference #566 AUTHCERTbot?
#566 = COA document verify. #568 = marketplace doubt fake copy product.
Going further
This week: sync FAKE-MAP verify_url bot, red team "confirm fake refund", measure fake_bot_fake_confirmed_invented. Activate page /authenticity embed.
Enzo
July 1, 2026
