E-commerce
July 1, 2026
"The bot takes a minute to reply, is that normal?" "I wrote, but the chatbot is no longer responding." "How much time before a human takes over after the bot?" Three tickets where chatbot delay expectations lack a clear response.
The e-commerce chatbot response delay support covers AI latency, unavailability, conversation timeout, human handoff delay, and bot service hours. Distinct from human support response delay (#401): here, the customer is asking about the speed and availability of the chatbot, not the email agents or human chat SLA.
This guide #877 deploys policy CHATRT-SUP, flow CRT-1 to CRT-8, and matrix CHATRT-MAP. Partner of the future bot communication latency (#878).
Summary
Why do chatbot delays generate tickets?
LLM latency, API queue, maintenance, or slow handoff: the customer perceives a slow or silent bot. The agent confuses a chatbot complaint with human support delay #401 or promises a speed that the infrastructure does not guarantee.
Five typical chatbot delay frictions
Slow response: several seconds or minutes before the bot messages
No response: customer message with no bot reply
Bot unavailable: widget offline or error message
Slow handoff: long wait after requesting a human agent
Wait time expectation: wants to know how long the bot will take
DTC Retail Example
DTC fashion, 10k chat tickets/month. After CHATRT-MAP: chatrt_expectation_resolution_rate 87%, latency repetitions under 7 days -42%.
CHATRT #877 vs RESP #401, HANDOFF #12, FALLBACK #279 and bot #878
Five pending-delay content items, five distinct angles from the client side.
Quick matrix
#877 CHATRT: manage latency tickets unavailability timeout bot handoff
HANDOFF #12: rules when to transfer bot to human
CONTEXT #155: context transmission handoff distinct speed
FALLBACK #279: degraded mode Shopify data unavailable
Bot #878: bot messages latency outage widget side
#877 = why is the bot slow? #401 = why is the human agent taking time.
Promise #877
Policy CHATRT-SUP, tree CHATRT-GATE, 8 macros, bot latency matrix versus human SLA versus widget communication #878, KPI chatrt_expectation_resolution_rate.
Which chatrt_* typologies to classify?
Action-oriented classifier: slow_response ≠ no_response ≠ handoff_delay ≠ expected_time.
Eight CHATRT-MAP typologies
chatrt_slow_response: bot responds but too slowly
chatrt_no_response: no bot response after message
chatrt_unavailable: widget offline or bot disabled
chatrt_expected_time: asks about normal bot response time
chatrt_handoff_delay: long wait after escalation to human
chatrt_timeout: conversation expired or auto-closed
chatrt_peak_hours: perceived slowness during peak hours
chatrt_compare_human: expects immediate human speed from the bot
Policy CHATRT-SUP: agent rules and escalation
The CHATRT-SUP policy sets realistic expectations without promising an instant response or blaming the customer.
Six CHATRT-SUP rules
ACKNOWLEDGE-FIRST: CHATRT-ACKNOWLEDGE macro to validate latency frustration
Documented ETA: CHATRT-ETA bot and handoff delay according to registry
Explain latency: CHATRT-EXPLAIN-LATENCY factors without infra jargon
Alternative if down: CHATRT-ALTERNATIVE email form if unavailable
Distinct handoff: chatrt_handoff_delay human SLA referral #401 if applicable
Incident → ops: recurring chatrt_no_response technical escalation
Situation matrix (agent)
Normal latency: EXPLAIN-LATENCY + ETA
Bot down: UNAVAILABLE + ALTERNATIVE
Slow handoff: HANDOFF-ETA + handoff #401 if slow agent
Timeout: EXPLAIN timeout + reopen chat
Flow CRT-1 to CRT-8: standard resolution
Eight sequential steps, SLA P3 chatrt < 24 h response pending, escalate ops if no_response is recurring.
Flow CRT-1 to CRT-8
CRT-1 Triage: read complaint, tag chatrt_*, bot or human #401?
CRT-2 Lookup: conversation logs, measured latency, bot incident status
CRT-3 Educate: ETA if expected_time, distinguish bot versus agent
CRT-4 Classify: chatrt_* via CHATRT-MAP
CRT-5 Execute: ACKNOWLEDGE, ETA, EXPLAIN, ALTERNATIVE, ESCALATE ops
CRT-6 Confirm: macro CHATRT-DONE waiting and next step
CRT-7 Test: customer understands normal delay or alternative
CRT-8 Close: KPI chatrt_expectation_resolution_rate
Eight ready-to-paste CHATRT-* macros
Macros aligned with chatbot delay log and internal handoff.
CHATRT-* Library
CHATRT-ACKNOWLEDGE: "We understand that waiting for the chatbot can be frustrating. Thank you for reporting this to us."
CHATRT-ETA: "Typical bot response time: {{délai}}. Human handoff: {{délai_handoff}} if requested."
CHATRT-EXPLAIN-LATENCY: "A delay of a few seconds is normal. Traffic peaks or complex questions can increase wait times."
CHATRT-UNAVAILABLE: "The chatbot is temporarily unavailable. Reason: {{statut}}."
CHATRT-ALTERNATIVE: "Please contact us via {{canal}} in the meantime: {{lien}}."
CHATRT-HANDOFF-ETA: "After requesting an agent, typical delay: {{délai}}. Conversation reference: {{id}}."
CHATRT-ESCALATE: "Issue escalated to the technical team. Update within {{SLA}}."
CHATRT-DONE: "Recap: {{situation}}. Normal wait time: {{délai}}. Alternative: {{canal}}."
CHATRT-GATE tree and agent-ready delay register
Decision tree before instant response promise or human SLA confusion.
CHATRT-GATE
Slow bot only? → EXPLAIN-LATENCY + ETA
Muted or down bot? → UNAVAILABLE + ALTERNATIVE + ESCALATE if incident
Slow human handoff? → HANDOFF-ETA + handoff RESP #401 if agent
Normal delay request? → ETA registry
After-sales service business issue? → resolve after-sales service, not CHATRT alone
Minimum internal registry
Document helpdesk: bot ETA P50 P95, handoff delay, bot service hours, alternative channel, incident procedure. Train agents: bot latency ≠ agent delay #401.
KPI, QA and handoff to bot #878
Measuring CHATRT detects over-promising on speed and bad human SLA silos.
Four CHART KPI
chatrt_expectation_resolution_rate : customer understands timeframe / total
chatrt_wrong_silo_rate : % wrongly routed to #401 solely for bot latency
chatrt_eta_given_rate : % of tickets with tracked ETA or EXPLAIN
chatrt_repeat_7d : same latency complaint within 7 days
Bot handof #878
Export CHATRT-MAP to outage latency widget messages. Guardrail CHATRT-WIDGET-ETA: display proactive wait before ticket frustration.
Edge cases: slow mobile, LLM timeout, closed hours
Three cases outside the standard flow.
Slow mobile connection
EXPLAIN-LATENCY client network possible. Do not blame infrastructure alone.
LLM or API timeout
UNAVAILABLE + ESCALATE ops. Immediate ALTERNATIVE.
Outside active bot hours
ETA service hours + ALTERNATIVE email. Distinction between bot off versus slow.
Agent training: 20 minutes CHATRT
Module: bot latency ≠ human SLA, systematic ACKNOWLEDGE, registry ETA, ALTERNATIVE if down, distinct HANDOFF-ETA.
Exercises
Ticket A: slow bot 30 s → EXPLAIN-LATENCY + ETA not empty excuse
Ticket B: bot stops responding → UNAVAILABLE + ALTERNATIVE + ESCALATE
Ticket C: slow human agent after bot → handoff RESP #401 not CHATRT alone
How Qstomy structures CHATRT in your stack
Qstomy route chatrt_*, displays P95 agent latency metrics and brief #878 for proactive widget messages.
Three building blocks
Routing: intent chatbot_latency vs human_response_time vs sav_issue
Latency dashboard: sync macros CHATRT-* ETA incident status
Bot #878: communicate widget-side outage latency
Scenario: DTC retail, Black Friday peak. Bot #878 proactive waiting message, CHATRT agents ETA. chatrt_expectation_resolution_rate goes from 58% to 85% in 3 weeks.
FAQ and CHATRT deployment checklist
FAQ
Should the bot respond instantly?
No. EXPLAIN-LATENCY of a few seconds is normal. Honest ETA.
Difference from #401?
#877 = chatbot speed. #401 = human support response delay.
Difference from #878?
#877 = agents managing waiting tickets. #878 = bot messages widget latency outage.
7-day Checklist
D1: CHATRT-SUP + CHATRT-MAP + bot handoff ETA registry
D2: 8 helpdesk macros
D3: routing matrix #401 #12 #279
D4: 20 min agent training
D5: chatrt_* tags + KPIs
D6: UNAVAILABLE vs EXPLAIN-LATENCY test
D7: bot brief #878 WIDGET-ETA-GATE
Interlinking
Enzo
July 1, 2026
