E-commerce
July 1, 2026
"I refuse to have my messages used for your AI." The bot responds "it is done" without a flag or blocks support. The customer opens a convopt_ ticket.
An e-commerce AI chatbot opt-out learning explains the scope, verifies the customer flag, routes to agent #909 if execution is required, and continues current support without breaking the UX.
This guide #910 covers bot_learnopt_* intents, flows LEARNbot LOB-1 to LOB-8, and LEARNOPT-RESPECT-GATE guardrails. Bot pair of the CONVOPT playbook (#909). Usecase: respecting customer choice on the widget side.
Summary
Why manage learning opt-out on the bot side?
Opt-out requests arrive mid-chat. Without a dedicated flow, the bot promises without an API flag or refuses to help. LEARNOPT-RESPECT-GATE reduces convopt_ #909 tickets and preserves trust.
What the tier 1 opt-out bot resolves
Explain scope: future past QA review map
Verify status: active opt-out flag registry
Initiate request: route to agent #909 if execution
Respect flag: current support without learning
Distinguish delete: opt-out ≠ deletion #903
Retail DTC Example
DTC, respect gate widget. learnopt_bot_deflect +41%, learnopt_bot_false_execute 0 in 4 weeks.
LEARNbot #910 vs CONVOPT #909, TRAINDATAbot #908, CONVDEL #903 and TRACKOPT #851
Seven privacy opt-out contents, seven distinct roles.
Quick matrix
#910 LEARNbot: explainer bot to verify and respect tier 1 learning opt-out
CONVOPT #909: agents execute confirm CO-5 flag
TRAINDATAbot #908: inform sources opt_out_route directs
TRAINDATA #907: agents traindata_opt_out handoff
CONVDEL #903: delete distinct opt-out transcript
TRACKOPT #851: opt-out marketing tracking distinct AI
Privacy #154: cross-functional data usage
Pipeline: widget learning refusal → #910 tier 1 → handoff #909 if EXECUTE is required.
Which bot_learnopt_* intents should be configured?
Eight opt-out intents mapped CONVOPT-MAP #909.
Eight bot_learnopt intents
bot_learnopt_request_route: initiate opt-out route #909
bot_learnopt_scope_answer: scope_copy perimeter map
bot_learnopt_status_check: flag_status CRM registry
bot_learnopt_respect_active: active flag behavior
bot_learnopt_vs_delete: reroute CONVDEL #903
bot_learnopt_vs_traindata: reroute TRAINDATAbot #908
bot_learnopt_still_used: agent route audit #909
bot_learnopt_feed_loop: consume LOG #909 scope_copy
Each interaction logs flag_checked intent convopt_* deflect_or_handoff.
How should CONVOPT-MAP #909 be consumed?
The bot reads CONVOPT-MAP #909 + bot fields: scope_copy, flag_status, opt_out_process, effective_delay, support_exception, delete_reroute, agent_handoff_url.
Bot opt-out guardrails
LEARNOPT-RESPECT-GATE: active flag = no dataset contribution
NO-FALSE-EXECUTE-BOT: do not promise EXECUTE without API flag
SCOPE-GROUNDED-BOT: scope_copy registry only
SUPPORT-CONTINUE-BOT: opt-out does not stop order support
DELETE-REROUTE-BOT: deletion → #903 CONVDEL
TRAINDATA-REROUTE-BOT: training sources → #908
AGENT-HANDOFF-BOT: EXECUTE identity → #909 agent
CONVOPT-FEED-LOOP-BOT: LOG #909 enriches scope_copy
LEARNBOT-SUP Policy in Six Rules
Six responsible opt-out rules on the widget side.
LEARNOPT-RESPECT-GATE: active flag respected session
NO-FALSE-EXECUTE-BOT: route #909 if technical execution
SCOPE-GROUNDED-BOT: scope from registry #909
SUPPORT-CONTINUE-BOT: opt-out ≠ refusal of customer support
STATUS-HONEST-BOT: flag status if unknown handoff
CONVOPT-FEED-LOOP-BOT: each LOG #909 widget review under 48 hours
Flow LEARNbot LOB-1 to LOB-8
Eight-step flow: incoming refusal classify flag scope answer route log handoff.
LOB-1 Ingest: opt-out message learning exclusion
LOB-2 Classify: learnopt_* vs traindata #908 vs delete #903
LOB-3 Flag check: LEARNOPT-RESPECT-GATE CRM lookup
LOB-4 Tier 1 answer: scope status respect copy
LOB-5 Route: request_route still_used agent_handoff
LOB-6 Feed loop: LOG #909 adjusts scope_copy widget
LOB-7 Handoff: EXECUTE identity → agent #909 context
LOB-8 Log: respect_rate deflect false_execute
TPL-LEARNOPT-SCOPE Example
“The opt-out covers: {{scope_copy}}. Your support remains active. Effective delay: {{effective_delay}}. More details: {{policy_link}}.”
TPL-LEARNOPT templates and touchpoints
Four short opt-out embed templates.
TPL-LEARNOPT-REQUEST
[opt_out_process map.] Identity verification. AGENT-HANDOFF if EXECUTE required.
TPL-LEARNOPT-SCOPE
[scope_copy map.] [support_exception map.] SCOPE-GROUNDED.
TPL-LEARNOPT-STATUS
[flag_status map.] If unknown: handoff #909. STATUS-HONEST.
TPL-LEARNOPT-RESPECT
[respect_copy map.] Active flag: no dataset contribution. SUPPORT-CONTINUE.
Touchpoints
"I refuse training": request_route + scope
"Active opt-out?": status_check tier 1
"Delete the chat": vs_delete reroute #903
LOG CONVOPT #909: feed_loop scope_copy
Edge cases: guest, delayed flag, opt-in revocation
Five cases outside the standard flow.
Guest without account: request_route email verify handoff #909
Flag not yet propagated: STATUS-HONEST delay effective_delay
Customer revokes opt-out: handoff #909 documented opt-in
Question training sources: #908 TRAINDATAbot
Ticket convopt_ despite bot: feed_loop review scope gaps
Essential learnopt_bot KPIs
Five LEARNbot monitoring metrics and correlation #909.
learnopt_bot_deflect: opt-outs resolved without a convopt_ ticket
learnopt_bot_respect_rate: % active flag sessions respected
learnopt_bot_false_execute: EXECUTE promises without target API 0
learnopt_bot_handoff_rate: % request_route towards #909
learnopt_bot_scope_compliance: % responses aligned with scope_copy
Target: false_execute at zero and respect_rate above 95%.
LEARNbot Anti-patterns
Five common bot opt-out errors.
Promising without API flag: NO-FALSE-EXECUTE route #909
Blocking support: SUPPORT-CONTINUE mandatory
Improvised scope: SCOPE-GROUNDED registry only
Confusing delete: DELETE-REROUTE #903
Ignoring active flag: LEARNOPT-RESPECT-GATE session
LEARNbot with Qstomy
Qstomy on Shopify: CONVOPT-MAP sync #909, CRM lookup flag, templates scope status respect, agent handoff, KPI learnopt_bot dashboard.
Scenario: DTC, 4 tickets/month convopt. Gate tier 1 respect. learnopt_bot_deflect +41%, convopt_ tickets -38% in 4 weeks.
Explore AI support and request a demo.
Checklist, FAQ and going further
LEARNbot Checklist (8 steps)
Sync CONVOPT-MAP #909: scope flag delay exception
Policy LEARNBOT-SUP: 6 RESPECT-GATE NO-FALSE-EXECUTE rules
8 intents bot_learnopt_*: flow LOB-1 to LOB-8
4 templates TPL-LEARNOPT-*: REQUEST SCOPE STATUS RESPECT
policy_link: public conversation opt-out page
agent_handoff: route #909 EXECUTE configured
Red team opt-out: false execute test flag respect
KPI Dashboard: learnopt_bot_* section 9 + delta convopt_
FAQ
Difference #909?
#909 = agents EXECUTE CONFIRM flag. #910 = bot tier 1 scope status respect.
Difference #908?
#908 informs training sources. #910 manages learning opt-out.
Can the bot activate the opt-out by itself?
No if VERIFY EXECUTE is required. NO-FALSE-EXECUTE + handoff #909.
Does opt-out block support?
No. SUPPORT-CONTINUE: after-sales service order help remains active.
Going further
This week: sync CONVOPT-MAP #909, templates scope status, agent handoff, measure learnopt_bot_respect_rate.
Enzo
July 1, 2026
