E-commerce
July 1, 2026
"The bot loads without saying anything." "Error message then total silence." "I wait for 2 minutes without knowing if it works." Three failures where the lack of latency communication worsens abandonment.
An AI chatbot with degraded e-commerce communication displays a proactive ETA, processing indicator, unavailability banner, and alternative channel, without promising immediate availability or leaving the customer in a vacuum.
This guide #878 covers intents bot_downtime_*, flow DOWNTIMEbot and KPI downtime_bot. Playbook bot pair CHATRT (#877). Use case: communicating latency and downtime on the widget side without frustrating.
Summary
Why communicate latency before frustration?
With no message for 10 seconds, the client closes the widget or opens a chatrt_ ticket. A calibrated bot displays typing_copy and eta_banner_copy as soon as the latency threshold is reached, without any NO-FAKE-READY violation.
What widget communication solves
Proactive ETA: eta_copy a few seconds normal map
Active indicator: typing_copy processing in progress map
Degraded banner: degraded_copy latency or incident map
Alternative channel: alternative_copy email form map
Ticket reduction: fewer chatrt_slow_response chatrt_no_response
DOWNTIMEbot vs CHATRT #877, FALLBACK #279, HANDOFF #12 and INCIDENT #278
Six bot reliability contents, six distinct angles.
Quick Matrix
#878 DOWNTIMEbot: widget messages, alternative banner ETA, outage latency
CHATRT #877: agents manage queue tickets, CRT-5
FALLBACK #279: degraded mode Shopify data API down
HANDOFF #12: human handoff rules, distinct queue communication
INCIDENT #278: incident communication, human macros, status page
HOURS #402: hours, bot off, distinct, slow
Pipeline: #878 widget warn → #877 agents if ticket anyway → #279 if API data down.
Which bot_downtime_* intents should be configured?
Eight degraded communication intents mapped to chatrt_* typologies #877.
Eight bot_downtime intents
bot_downtime_eta: eta_copy normal response time map
bot_downtime_typing: typing_copy processing in progress map
bot_downtime_slow: slow_copy prolonged latency map
bot_downtime_unavailable: unavailable_copy bot offline map
bot_downtime_degraded: degraded_copy degraded mode map
bot_downtime_alternative: alternative_copy backup channel map
bot_downtime_handoff_wait: handoff_wait_copy waiting for agent map
bot_downtime_recovery: recovery_copy service restored map
Resolve bot infra incident → ops #279 FALLBACK. Widget communicates only.
How to consume CHATRT-MAP #877?
The bot reads CHATRT-MAP #877 + widget fields: eta_copy, typing_copy, slow_copy, unavailable_copy, degraded_copy, alternative_copy, handoff_wait_copy, recovery_copy, latency_threshold_ms.
Latency communication guardrails
PROACTIVE-ETA-BOT: eta_copy if latency > threshold before response
TYPING-INDICATOR-BOT: typing_copy as soon as customer message is sent
NO-FAKE-READY-BOT: do not display ready if LLM is not ready
ALTERNATIVE-OFFER-BOT: alternative_copy if unavailable or slow > max
DEGRADED-BANNER-BOT: degraded_copy if KB_ONLY mode #279
NO-BLAME-USER-BOT: never blame the customer for slowness
FALLBACK-REROUTE-BOT: Shopify API down → #279 degraded modes
DOWNTIMEBOT-SUP Policy in six rules
Six rules for responsible degraded communication.
PROACTIVE-ETA-BOT: waiting message before 8 s of silence
TYPING-INDICATOR-BOT: indicator visible during generation
NO-FAKE-READY-BOT: widget open only if the bot is actually active
ALTERNATIVE-OFFER-BOT: email or form if down > 30 s
DEGRADED-BANNER-BOT: honest banner if degraded mode is active
RECOVERY-NOTIFY-BOT: recovery_copy when service is restored
Flow DOWNTIMEbot DTB-1 to DTB-8
Eight-step communication flow latency embed widget message opening sending.
DTB-1 Health check: active bot? if not, immediate unavailable_copy
DTB-2 Open widget: NO-FAKE-READY if maintenance flag true
DTB-3 On message send: TYPING-INDICATOR triggered
DTB-4 Latency watch: if > threshold PROACTIVE-ETA eta_copy or slow_copy
DTB-5 Mode check: degraded_copy if FALLBACK #279 KB_ONLY HANDOFF_ONLY
DTB-6 Respond or fallback: normal response or alternative_copy
DTB-7 Handoff wait: handoff_wait_copy if human escalation
DTB-8 Log: latency_ms eta_shown alternative_offered recovery_sent
Example TPL-DOWNTIMEbot-ETA
« [typing_copy map.] [eta_copy map.] PROACTIVE-ETA-BOT. »
TPL-DOWNTIMEbot and touchpoint templates
Four short templates for degraded communication embed.
TPL-DOWNTIMEbot-ETA
[typing_copy map.] [eta_copy map.] TYPING-INDICATOR-BOT.
TPL-DOWNTIMEbot-SLOW
[slow_copy map.] [alternative_copy map optional.] PROACTIVE-ETA-BOT.
TPL-DOWNTIMEbot-UNAVAILABLE
[unavailable_copy map.] [alternative_copy map.] ALTERNATIVE-OFFER-BOT.
TPL-DOWNTIMEbot-DEGRADED
[degraded_copy map.] [handoff_wait_copy map optional.] DEGRADED-BANNER-BOT.
Touchpoints
Widget Opening: health check unavailable if down
Sending message: typing + ETA threshold 3-8 s
Latency > 15 s: slow_copy + proactive alternative
FALLBACK Mode #279: degraded_banner top widget
Edge cases and reroutes
Five cases outside of standard communication.
Shopify API timeout: FALLBACK #279 degraded + handoff
Platform Incident SEV1: unavailable + link status #278
Outside bot hours: HOURS #402 reroute
Slow human handoff: handoff_wait_copy + CHATRT #877 if ticket
Customer asks why it's slow: slow_copy + offer handoff #12
Essential downtime_bot KPIs
Five DOWNTIMEbot management metrics.
downtime_bot_frustration_deflect: closed without ticket chatrt_ repeat 7d
downtime_bot_eta_shown_rate: % latency > threshold with eta_copy
downtime_bot_fake_ready_violations: widget open bot down target 0
downtime_bot_alternative_offer_rate: % unavailable slow with alternative_copy
downtime_bot_abandon_after_slow: % widget closure after slow without response
Target: fake_ready_violations 0 and abandon_after_slow decreasing after PROACTIVE-ETA.
DOWNTIMEbot anti-patterns
Five common latency communication errors.
15s silence without a message: PROACTIVE-ETA typing mandatory
Widget open bot maintenance: NO-FAKE-READY immediate unavailable
Promising an instant response: honest eta_copy a few seconds
No alternative if down: ALTERNATIVE-OFFER email form
Hiding degraded mode: DEGRADED-BANNER if FALLBACK #279 active
DOWNTIMEbot with Qstomy
Qstomy on Shopify: latency watch P95, CHATRT-MAP sync eta alternative, FALLBACK #279 banner auto, KPI downtime_bot dashboard.
Scenario: retail Black Friday, P95 latency 12 s. PROACTIVE-ETA + typing, chatrt_ tickets -38%. downtime_bot_frustration_deflect 79% in 2 weeks.
Explore AI support and request a demo.
Checklist, FAQ and going further
DOWNTIMEbot Checklist (8 steps)
Sync CHATRT-MAP #877: eta alternative handoff_wait copies widget
Policy DOWNTIMEBOT-SUP: 6 rules PROACTIVE-ETA NO-FAKE-READY ALTERNATIVE
8 intents bot_downtime_*: flow DTB-1 to DTB-8
4 templates TPL-DOWNTIMEbot-*: ETA SLOW UNAVAILABLE DEGRADED
Latency threshold 3-8 s: configurable eta_copy trigger
Link FALLBACK #279: degraded_banner if KB_ONLY HANDOFF_ONLY
Red team latency: silence fake_ready no alternative test
KPI Dashboard: downtime_bot_* section 9 fake_ready abandon
FAQ
Difference #877?
#877 = agents answer waiting tickets. #878 = widget communicate before ticket.
Difference #279?
#878 = user messages latency breakdown. #279 = technical behavior Shopify API down.
What delay to announce?
eta_copy registry #877. A few seconds is normal, more if peak.
Bot completely down?
unavailable_copy + alternative_copy. NO-FAKE-READY close or clear message.
Going further
Customer Support chatbot delay (#877)
This week: activate PROACTIVE-ETA threshold 5 s, test unavailable alternative flow, sync degraded_banner FALLBACK #279 test.
Enzo
July 1, 2026
