E-commerce
June 30, 2026
Your homepage promises "express delivery", "frictionless returns", "traced ingredients". The customer opens a delayed package ticket: "Sorry for the inconvenience, your request will be processed within 5 business days." The promise comes from marketing. The response comes from a bank template.
Butterflai points out that an inconsistent voice creates friction at moments when trust is at stake: checkout, error, complaint (Butterflai, e-commerce tone of voice). FitGap believes that the gap between promise and frontline execution is bridged by playbooks and QA loops, not by more generic training (FitGap, brand playbooks 2026).
This guide #297 covers human support aligned with brand promise: tone, proofs, macros. It complements bot voice (#125) and precedes promise bot (#298) focusing on the angle of human customer service + cross-channel consistency.
Summary
Why does support either reinforce or degrade the brand promise?
Support is the post-promise moment of truth. The customer compares what they read on the site with what they hear from the agent or read in the email.
Three common gaps
Tone: warm website, cold or corporate support
Proof: marketing claim without concrete proof in the response
Policy gap: PDP promise u2260 actual policy applied by the agent
Business impact
DTC Playbook notes that customers with issues that are well resolved sometimes become the best promoters: recovery builds trust (DTC Playbook, support). Lexsis estimates that brands with voice-aligned support achieve ~92% of human CSAT in automation trained on the best conversations (Lexsis, brand voice 2026).
DTC Example
"Clean" skincare brand: homepage promises INCI transparency. Damaged package tickets used to be answered with logistics jargon. After BRAND-DAMAGE macros + batch proof + empathetic tone: complaint CSAT +0.7 pt, chargebacks u221218%.
How does it differ from the bot voice (#125) and the AI guide (#298)?
Three neighboring contents, three perimeters.
Bot voice (#125)
Bot voice (#125): charter your chatbot, prompts, AI configuration. The #297: human agents, Gorgias macros, live QA.
Promise bot (#298)
Promise bot (#298): PROMISE-RULES, BRAND-PROMPT-01, AI guardrails. The #297 sets the common promise + evidence repository for both bot and human.
Generic templates
Support templates: response structure. The #297: brand claim alignment in each macro.
Response quality (#116)
Response quality (#116): factual accuracy. The #297 adds identity consistency and evidence.
Promise #297
Promise stack, situation taxonomy, BRAND-SUPPORT grid, macros, QA scorecard, gold standard, KPI drift.
How do you break down your brand promise for support?
Vague marketing promises ("premium quality") do not guide an agent. Break it down into an operational promise stack.
Four Layers
Main Claim: a website tagline (e.g., "Free 30-day returns")
Proof: certificates, processes, numbers, policy links
Limits: exceptions (personalized products, hygiene, sales)
Tone: 3 adjectives + 3 forbidden terms (e.g., warm, direct, expert / never condescending, never "dear valued customer")
PROMISE-STACK-01 Document
Notion table: Claim column | Agent proof | Linked macro | Source page. Every homepage claim must have a corresponding support row. Sameness alert: voice drift is cumulative, each "acceptable" response fragments the brand (Sameness, brand governance).
60-min Audit
Homepage + 3 best-selling PDPs + welcome email + 10 random tickets. Highlight claims without proof in agent responses.
What taxonomy of brand situations should be mapped?
Tag tickets with brand_ctx_* for targeted QA, not just intent support.
High-stakes brand situations
brand_ctx_promise_test: customer cites website claim ("you promise...")brand_ctx_trust: scam, quality, greenwashingbrand_ctx_recovery: damaged parcel, delay, wrong itembrand_ctx_vip: loyal customer, expected relational tonebrand_ctx_public: threat of Google review, social mediabrand_ctx_values: ethics, sustainability, origin
QA Prioritization
Brand scorecard on 100% of brand_ctx_promise_test and brand_ctx_public, 10% sample of others. FitGap recommends 6 to 12 scorecard items with "what good looks like" examples (FitGap).
Which BRAND-SUPPORT grid should be applied to each response?
Before sending a macro or free response, the agent applies the BRAND-SUPPORT-01 grid (30 seconds).
5 criteria
Recognition: customer emotion named in 1 sentence
Aligned claim: faithful policy reformulation, no over-promising
Proof: link, batch number, quantified delay, process screenshot
Tone: formal/informal address matching the style guide, length adapted to the channel
Next step: clear action + owner deadline
Package recovery example
Bad: "Your ticket is in progress."
Good: "I understand the disappointment, especially for a gift. Our reinforced packaging promise applies: I am sending you the replacement today, without returning the damaged package. Tracking within 2 hours."
Brand prohibitions
Blacklist: "dear customer", "we inform you that", "in accordance with our procedures" without concrete next steps, minimization ("it's only a 2-day delay"). Butterflai: clarity first on errors and support (Butterflai).
Which BRAND-* macros should be aligned with the promise for each ticket type?
Ecommerce Circle recommends ~25 top ticket macros, reviewable in 10 s of personalization (Ecommerce Circle, macros 2026). The #297 injects the promise into each macro.
BRAND Library (extracts)
BRAND-WISMO: status + ETA + delivery claim reminder if delayed
BRAND-RETURN: return process + deadline + branded policy link
BRAND-DAMAGE: empathy + friction-free replacement + proof of packaging
BRAND-TRUST: proof of cert/reviews + invitation to transparency
BRAND-VALUES: sourcing, sustainability with link to proof page
BRAND-PROMISE-CITE: customer quotes homepage, claim answer + proof + limitation
Standard BRAND-PROMISE-CITE Macro
“You are right to recall our promise “ {{claim_homepage}} ”. Here is how it applies to your order #{{order}}: {{application_concrete}}. Full details: {{policy_url}}. I remain available if any point remains unclear.”
Personalization Rule
Each macro: 1 line to edit (first name, order details, emotional context). Macros not personalizable in 10 s = to be rewritten.
How to install a QA brand loop on conversations?
Brand QA transforms the charter into measurable behaviors, not a forgotten PDF.
8-item Scorecard (example)
Emotion recognition (0/1)
Charter tone (0/1)
Accurate claim/policy (0/1)
Proof provided (0/1)
No over-promising (0/1)
Next step + deadline (0/1)
Visible personalization (0/1)
Blacklisted items absent (0/1)
FitGap Pace
Weekly sample: 15 tickets brand_ctx_*. Rating + snippet evidence. Gap → 7-day coaching task + Guru/Notion macro update. Recurring gap 5+ times → dedicated "moment card" playbook.
Gold standard
Lexsis: collect 50 best CSAT 5/5 conversations as a voice reference (Lexsis). Integrate them into agent onboarding and bot training #298.
How to train agents and synchronize marketing, support, and product?
Support cannot keep a promise that the product or the website contradicts.
Agent Onboarding (D1-D3)
D1: read PROMISE-STACK-01 + 10 gold standards
D2: shadow + rewrite 5 tickets with BRAND grid
D3: supervised answers, scorecard ≥ 7/8
Full D1-D90 program: see real case training (#299).
Monthly Cross-Functional Review (45 min)
Support: top 5 brand_ctx_promise_test. Marketing: claim to be clarified on site. Product: PDP patch if recurring objection. Link personas #296 for customer language.
Slack Channel #brand-support
Ambiguous ticket → screenshot + question "how do we keep the promise here?" Answer from marketing or ops owner within 24 hours. Document decision in PROMISE-STACK.
Which KPIs should be used to measure the consistency between the promise and the support?
Measure perceived consistency, not just overall CSAT.
Monthly KPIs
Brand QA score: average scorecard for 8 items
Promise test rate: brand_ctx_promise_test tickets / total
Recovery CSAT: CSAT for brand_ctx_recovery tickets
Chargeback / 1★ review post-complaint
Macro drift: macros modified outside the process (Gorgias audit)
Bot vs human CSAT gap: difference < 3 target Lexsis points
Alert signal
If promise_test rate increases without a website patch: marketing is over-promising or support is under-informing. Joint action within 2 weeks.
How does Qstomy maintain alignment promise on support?
Qstomy applies PROMISE-STACK and gold standard to bot responses and human handoff briefings.
Capabilities
Brand voice profile + documented claims. BRAND-* macros injected. Auto tone scorecard on bot responses. Agent handoff: cited claim, suggested proof, tone notes. Monthly QA export brand_ctx. Alignment voice #125.
Quantified DTC Scenario
Premium mode, 2,800 tickets/month, brand QA score 5.8/8, 34 tickets/month "you promise on the site." PROMISE-STACK deployment + 6 BRAND macros + weekly QA. After 10 weeks: brand QA 7.4/8, recovery CSAT +0.9 pt, promise_test tickets −31% (site clarification), bot/human CSAT gap 1.8 pt.
See AI support, Shopify, demo.
Which playbooks can be used to align support with brand promise?
Playbook 1: promise stack (3 h)
Audit homepage + PDP + 10 tickets. Complete PROMISE-STACK-01: claim, proof, limit, macro, URL.
Playbook 2: BRAND macros (4 h)
Rewrite top 10 Gorgias macros using the section 5 grid. 10 s personalization test. Publish forbidden list.
Playbook 3: QA scorecard (2 h)
Configure 8 FitGap items. Sample 15 tickets/week. Coaching gap template.
Playbook 4: gold standard (2 h)
Extract 20 CSAT 5/5 conversations. Annotate what makes them on-brand. Day 1 onboarding.
Playbook 5: cross-functional review (45 min/month)
Support + marketing + product: top promise_test, 2 website actions, 1 macro patch.
Useful linking
This week: open 10 random tickets from the month. How many cite proof linked to a homepage claim? If fewer than 4, draft BRAND-PROMISE-CITE before any other macro project.
Enzo
June 30, 2026
