E-commerce
July 1, 2026
"Is this a robot or a person?" "Who is supervising your AI?" "I want to speak to a human, not a machine." Three tickets where the human transparency behind chatbot lacks a clear answer.
The e-commerce supervision team support questions explains who answers, when an agent takes over, how the bot is supervised, and who is responsible, distinct from the technical handoff (#12) and the AI disclosure (#860).
This guide #915 deploys policy HUMANBOT-SUP, flow HB-1 to HB-8, and matrix HUMANBOT-MAP. Future customer service pair bot supervision (#916).
Summary
Why does the human behind the bot generate tickets?
Customer suspicious of AI, wants a real person or to understand who validates the bot's answers. Agent replies "it's automated" without explaining supervision or the human path. Without HUMANBOT-MAP, this causes confusion with handoff_ #12 or aitran_ #860.
Five typical frictions regarding humans behind the bot
Robot or human: does not know who is answering right now
Unclear supervision: do human agents verify the bot's outputs
Responsibility: who is responsible if the bot makes a mistake
Human access: how to reach an agent
Team trust: where does the support team work
DTC Retail Example
DTC Fashion, 6 humanbot_ tickets/month. After HUMANBOT-MAP: humanbot_trust_resolution_rate 88%, repeated requests for human agents -36%.
HUMANBOT #915 vs HANDOFF #12, AITRAN #860, Governance #142 and bot #916
Seven trust-bot contents, seven distinct angles.
Quick matrix
#915 HUMANBOT: explain supervision team human responsibility behind bot
HANDOFF #12: rules for when to transfer bot to human
AITRAN #860: AI status disclosure separate team response
Governance #142: internal rules RACI validation
OFFAIRESP #913: offensive response apology incident
TRAINDATA #907: training data separate supervision
Bot #916: explain role of human agents widget
#860 = is this response AI? #915 = who supervises and who is responsible.
Promise #915
Policy HUMANBOT-SUP, tree HUMANBOT-GATE, 8 macros, supervision team registry, KPI humanbot_trust_resolution_rate.
Which humanbot_* typologies should be classified?
Action-oriented classifier: identity u2260 supervision u2260 responsibility u2260 access.
Eight HUMANBOT-MAP typologies
humanbot_robot_or_human: who is responding right now
humanbot_supervision_ask: are agents checking the bot
humanbot_team_identity: who makes up the support team
humanbot_responsibility_ask: responsibility for bot errors
humanbot_reach_human: how to speak to an advisor
humanbot_always_bot_fear: fear of never reaching a human
humanbot_hours_human: availability of human agents
humanbot_trust_concern: general mistrust of AI alone
HUMANBOT-SUP Policy: agent rules and team registry
The HUMANBOT-SUP policy establishes honest transparency without over-promising permanent human presence.
Six HUMANBOT-SUP rules
HONEST-DISCLOSURE: state bot vs human if known
SUPERVISION-EXPLICIT: supervision_copy from register
REACH-HUMAN-CLEAR: documented handoff human path
RESPONSIBILITY-MARQUE: brand responsible, not bot alone
Do not deny team: honesty of existing agents
Incident → #913: serious handoff error OFFAIRESP
Minimum supervision team register
Model: bot first human on demand or hybrid
Supervision: QA review escalation validation
Human hours: live agent slots
Handoff: keyword SLA button
Responsibility: brand supervises bot
Flow HB-1 to HB-8: handling human-behind-the-bot questions
Eight sequential steps, humanbot P3 SLA < 48 h, live handoff if reach_human is urgent.
Flow HB-1 to HB-8
HB-1 Triage: supervision team vs handoff #12 vs disclosure #860?
HB-2 Classify: humanbot_* via HUMANBOT-MAP
HB-3 Registry: lookup supervision team hours handoff
HB-4 Explain: honest registry HUMANBOT macro
HB-5 Reach: REACH-HUMAN if agent requested
HB-6 Reassure: trust concern always_bot_fear
HB-7 Escalate: error incident → #913 if applicable
HB-8 Close: humanbot_trust_resolution_rate KPI + brief #916
Eight HUMANBOT-* macros ready to paste
Aligned macros disclosure supervision reach responsibility.
HUMANBOT-* Library
HUMANBOT-DISCLOSURE: "You are chatting with {{canal}}. Agents available: {{heures}}."
HUMANBOT-SUPERVISION: "Our chatbot is supervised by {{équipe}}. Quality review: {{process}}."
HUMANBOT-TEAM: "Support team {{marque}}: {{description}}."
HUMANBOT-RESPONSIBILITY: "{{marque}} remains responsible. Bot error: we will correct it."
HUMANBOT-REACH: "Human advisor: {{chemin}}. Delay: {{SLA}}."
HUMANBOT-REASSURE: "An agent can take over at any time via {{chemin}}."
HUMANBOT-HOURS: "Live agents: {{créneaux}}. Outside these hours: {{alternative}}."
HUMANBOT-DONE: "Summary: {{question}}. Answer: {{résolution}}. Reference: {{id}}."
HUMANBOT-GATE tree and immediate human reach
Decision tree before answering "everything is automated" or delaying handoff.
HUMANBOT-GATE
Technical handoff request only? → handoff #12 rules
AI response status disclosure? → handoff AITRAN #860
Team supervision question? → SUPERVISION TEAM
Wants a human now? → REACH live handoff if business hours
Fear of never getting a human? → REASSURE + REACH
Serious bot error? → handoff OFFAIRESP #913
Reach human urgent
humanbot_reach_human during live hours: immediate handoff or ticket takeover. Outside hours: HUMANBOT-HOURS + email callback if documented.
KPI, QA and hand-off to bot #916
Measuring HUMANBOT detects team denial and opaque human reach.
Four HUMANBOT KPIs
humanbot_trust_resolution_rate: trust questions resolved without escalation
humanbot_registry_compliance: % of answers aligned with team registry
humanbot_reach_human_success: % of reach_human with successful handoff
humanbot_deny_team_rate: denial of existing agents, target is low
Handoff #916
Export HUMANBOT-MAP to bot: humanbot_robot_or_human humanbot_supervision_ask priority. Guardrail HUMANBOT-REGISTRY-GATE brief #916 supervision_copy widget.
Edge cases: agent presents as human, outsourced team, B2B
Three cases outside the standard flow.
Agent says "I am the bot" by mistake
Correct DISCLOSURE. Train agents on consistent HUMANBOT-DISCLOSURE macro.
Outsourced BPO support
Honest TEAM copy: partner under brand supervision if the register indicates so.
B2B client demands an account manager
REACH routes to a distinct B2B segment if documented. Do not promise an AM if nonexistent.
Agent training: 20 minutes HUMANBOT
Module: DISCLOSURE SUPERVISION REACH, distinguish #12 #860 #916.
Exercises
Ticket A: "is it a robot?" → DISCLOSURE SUPERVISION
Ticket B: "I want a human" → REACH handoff #12
Ticket C: "who is responsible?" → RESPONSIBILITY brand
How Qstomy structures HUMANBOT in your stack
Qstomy route humanbot_*, sync supervision team registry, DISCLOSURE REACH macros and handoff #916 registry gate.
Three building blocks
Routing: intent human_behind vs handoff vs aitran
Team registry: supervision hours model handoff
Bot #916: tier 1 supervision agents widget
Scenario: DTC, 6 tickets/month humanbot. REGISTRY-FIRST agents, bot #916 tier 1. humanbot_trust_resolution_rate goes from 70% to 89% in 5 weeks.
FAQ and deployment checklist for HUMANBOT
FAQ
Promise human 24/7?
No if hours are limited. HUMANBOT-HOURS is honest.
Difference with #12?
#12 = when to handoff. #915 = who supervises and who is responsible.
Difference with #860?
#860 = AI response status. #915 = human team behind the bot.
Difference with #916?
#915 = ticket agents. #916 = bot explaining supervision widget.
7-day Checklist
D1: HUMANBOT-SUP + HUMANBOT-MAP + team registry
D2: 8 helpdesk macros
D3: routing matrix #12 #860 #142
D4: 20-min agent training DISCLOSURE REACH
D5: humanbot_* tags + KPIs
D6: test supervision vs reach vs responsibility
D7: bot brief #916 REGISTRY-GATE
Interlinking
Enzo
July 1, 2026
