You have worked the five modules — the foundation, the queue engine, quality & the team, measure & the customer signal, and the operating system. Each one ended with a graded artifact: the voice guide, the policy table, and the full macro library in English and Arabic; the ticket cluster map, the bug escalation report, and five help articles; the QA rubric, the scored batch report, and the onboarding plan for a new hire; the CSAT driver analysis, the VoC report, and the three-lever action plan; the weekly rhythm checklist, the incident comms kit, and the product handoff. The capstone is where those five stop being five.
This is the integrative project that earns the credential. One brand, one support operation, end to end — a complete system where the voice guide is still load-bearing in the incident reply, the macros trace to the same policy table that governs the escalation, and the VoC report closes the loop back to the foundation. That is the thing “Certified Customer Support with Claude” attests, and it is a higher bar than five modules completed: a certified support operator is not someone who watched the track, it is someone who can run a complete support operation where every layer coheres and every reply that leaves the building has a named human behind it.
Five modules teach you the playbooks. The capstone proves you can run them as one system — a complete support operation where voice, policy, macros, the queue, the QA standard, the measure loop, and the operating rhythm all inherit the same source of truth. A reviewer should not be able to tell that the macro library, the CSAT analysis, and the March incident reply were drafted in different sessions for different problems. Integration is the grade, not five separate artifacts that each look fine alone.
The brief
Produce a complete support system for one brand, and prove it by running that team’s operation from voice guide to operating rhythm, end to end. We recommend your own team — the capstone then doubles as deployable infrastructure you keep and run (see Bring your own team) — or the sample, Mizan’s support team, the GCC bookkeeping SaaS whose CX operation is threaded through the whole track, if you’d rather learn on neutral ground.
You already know the Desktop motion: open the team’s folder in Claude Desktop, approve each read in the “Ask permissions” prompt, and assemble the system in the chat and the file pane. No terminal on this path. Claude drafts fast; you own every reply that leaves the building. A person reviews every macro before it goes into rotation, clears every escalation above the Solo threshold, and signs off on every incident communication before it reaches a customer.
Capstone deliverable — Support-in-a-Box (one team, one brand, end to end)
1. The foundation (Foundation / M1)
- support-voice.md — how you sound, apology pattern, words use/avoid
- support-policy.md — 3-tier authority table, escalation line, never-say list
- macros.md — 10 on-voice templates with [placeholders]
- support-voice-ar.md — Arabic register, authored not translated (bilingual teams)
2. The queue engine (Queue Engine / M2)
- ticket cluster map — top 10 issue types with counts from a real export
- bug escalation report — paste-ready for the tracker, one bug fully surfaced
- 5 help articles — for the top deflectable questions by volume
3. Quality & the team (Quality / M3)
- QA rubric — 4–5 criteria anchored to support-voice.md + support-policy.md
- scored batch report — 20 replies, team-wide findings, named patterns
- onboarding.md — 5-day ramp for a new hire, milestones and cheat sheet
4. Measure & the customer signal (Measure / M4)
- CSAT driver analysis — sample bias named, drivers split process vs reply
- VoC report — 6–8 themes with counts and impact rating
- action plan — 3 levers for product and leadership, each with an owner
5. The operating system (OS / M5)
- weekly rhythm checklist — 5 steps, each with owner + skip consequence
- incident comms kit — 6 pieces, all templated, bilingual
- product handoff — 1-page VoC summary for the product team
Every piece must inherit #1. That inheritance is the whole assignment.The integration is not a formality. A reviewer will read the incident reply and check whether it sounds like the same team that wrote the macros. They will read the VoC report and check whether the agent warmth theme matches what the QA rubric grades. They will read the Arabic incident comms and check whether the register was authored or translated. Where those seams hold, the system works. Where they drift, the capstone flags a revise.
How it’s assessed — the master rubric
Five criteria, each scored meets / nearly / not yet. One “nearly” anywhere is a revise, not a pass — because a support system that coheres in four places and drifts in the fifth still means a customer in crisis gets a reply that sounds like it came from a different team, and that is exactly when coherence matters most. The bar throughout is a single question: would Layla Al-Nasser — or the CX lead you respect most — send this reply, post this help article, and stand behind this escalation?
Master rubric — Support-in-a-Box
Each criterion: meets / nearly / not yet. The bar: would the CX lead send
this reply, post this article, and stand behind this escalation?
1. Brand consistency
meets The VoC report reads like the same team that wrote the macros.
The incident reply follows the same voice as the macro library.
A reader can't tell which pieces were drafted in different
sessions or which language came first.
nearly Mostly coherent, but one piece drifts in tone — a macro that's
more formal than the voice guide, or an incident reply that
opens with "We apologize for any inconvenience."
not yet Different pieces sound like different teams. The voice guide
says own the mistake in the first sentence; the macros bury it.
2. Human decision gates
meets Every escalation above AED 150, every annual-plan change, every
incident reply has a named human owner and a named approval step.
The Solo / Manager / Never tiers are enforced — no Solo-tier
agent committed a Manager-tier promise.
nearly Most gates held, but one promise slipped through at the wrong
authority level, or an escalation is documented without a named
approver.
not yet No authority structure visible. Agents make refund decisions
without a tier; incident replies go out without sign-off.
3. Data safety
meets Customer PII — names, emails, account numbers — was scrubbed
before every analysis task. Patterns only, never personal data,
entered the chat. The cluster map uses ticket counts, not
customer identifiers.
nearly One analysis contains a real customer name or email that should
have been replaced with a [placeholder].
not yet Customer data handled without controls. The CSAT analysis or VoC
report contains identifiable customer information.
4. Arabic quality
meets The MENA team's Arabic replies are authored in register, not
translated. The Arabic incident comms ship simultaneously with
English. The never-say list applies in Arabic too — no
نعتذر عن أي إزعاج قد يكون حدث.
nearly Competent but translation-flavored, or the simultaneous bilingual
requirement was skipped for the incident kit.
not yet English run through translation. Reads foreign. Or the Arabic
pieces were omitted.
5. Completeness
meets All five phases present and coherent. The macros inherit the voice
guide. The onboarding inherits the QA rubric and the cheat sheet.
The operating system closes the loop back to M1. A new hire
could be onboarded and a real incident managed from this kit alone.
nearly A phase is present but disconnected — a QA rubric that doesn't
reference the voice guide, or an incident kit that has no template
for the customer-facing update.
not yet Fewer than five phases, or phases present but not connected to
each other.Notice that three of the five criteria grade the seams between phases, not the quality of any single document. The modules already graded each artifact alone; the capstone grades whether the voice guide, the QA rubric, the CSAT analysis, and the incident reply hold together as one operation.
The five gates — what each module proved
The foundation (M1) proved that you can anchor an entire support operation to two documents and a macro library. The reviewer checks that support-voice.md has a real apology pattern and a use/avoid word list — not vague direction like “be warm” — and that every macro in macros.md can be traced to a specific voice rule and a specific policy tier. The Arabic voice guide is authored, not a translation of the English.
The queue engine (M2) proved that you can turn a raw ticket export into a prioritised action plan. The reviewer checks that the cluster map is structured by volume and deflectability — not alphabetical or by gut — and that the bug escalation report names the pattern (annual-plan customers, all after March 1) cleanly enough to paste into an engineering tracker without a translation call.
Quality & the team (M3) proved that you can set a standard and hold it. The reviewer checks that the QA rubric criteria are directly traceable to support-voice.md (not invented from scratch), that the scored batch report names team-wide patterns rather than just individual scores, and that onboarding.md would genuinely prepare a new hire — like Hani Al-Rashidi in his first support role — to pass the rubric within five days.
Measure & the customer signal (M4) proved that you can read a CSAT number honestly. The reviewer checks that the driver analysis names the self-selection bias (a response rate of 18% skews toward frustrated users), splits the drivers into process issues versus reply issues rather than blaming agents as a single bucket, and produces an action plan with levers that product and leadership can actually pull.
The operating system (M5) proved that you can close the loop. The reviewer checks that the weekly rhythm checklist has named owners and named skip consequences — not just a checklist — and that the incident comms kit covers all six pieces: the internal flag, the customer-facing update, the interim reply, the resolution note, the billing fix communication, and the post-mortem summary.
Bring your own team
Use your own team. We recommend it for one blunt reason: the capstone stops being coursework and becomes two things at once — the actual support infrastructure your team runs on, and the credential that proves you built it to a professional standard. The voice guide, the macro library, the QA rubric, the CSAT loop, the incident comms kit — this is not a portfolio artifact, it is the system you open on Monday morning. That is day-one ROI: the assessment and the operational work are the same artifact.
The sample — Mizan’s support team — is here for when neutral ground is the point: a support lead who wants to practice the integration once before rolling it out on live customer tickets, or a team buyer who wants a worked reference to show the team what the output looks like before committing to the build. Either path clears the same rubric.
Sample brand: Mizan’s support team {#sample-brand-mizan}
Mizan is a GCC small-business bookkeeping SaaS with 9,000+ customers, a five-person support team led by Layla Al-Nasser, and a Q1 story that covers every module in the track. Use it as your sample and the full arc is already threaded for you:
The foundation built around Mizan’s actual voice — warm, plain, short sentences, owns the mistake in the first sentence — and the three-tier policy table where Solo agents can apply credits and comp one month for a clear Mizan error, Managers own refunds above AED 150 and any annual-plan change, and the never list includes admitting legal fault or inventing an order number engineering hasn’t confirmed.
The queue engine run on a real ticket export: 53 VAT export issues per week at the top, 47 login resets, 38 billing questions — and, buried at slot nine, 11 annual-plan double-charge tickets all dated on or after March 1, all from the same billing run. The bug escalation report that names the pattern cleanly enough for the engineering tracker.
The quality batch scored against the Mizan QA rubric, with Layla grading a week of replies and the scored report showing team-wide findings — including the two patterns that drove Q1 CSAT down. Hani Al-Rashidi onboarded in his first support role against a five-day plan that references the rubric and the cheat sheet.
The CSAT analysis that opens by naming the bias — 18% response rate, frustrated users over-represented — then splits the drivers: slow first-touch on VAT export issues (a process issue, fixable with a help article and a triage macro) versus agents over-hedging on timelines (a reply issue, fixable with a macro that sets a specific date and owns it). The positive driver: warmth and ownership when an agent names the mistake first.
The operating system built for Layla’s weekly rhythm and the March incident — 47 annual-plan customers locked out after the billing run, a three-hour outage window, incident comms shipping simultaneously in English and Arabic from templates that follow the same voice guide the macro library does.
Run all five phases on Mizan and the capstone Q1 story is complete. Then run it again on your own brand, and the infrastructure is yours.
What the certificate honestly means today
Clear the rubric and you earn “Certified Customer Support with Claude” — a verifiable page at /cert/[id] you can share and add to LinkedIn. It is the credential of the team track: proof that you can run a complete support operation on a coherent system, not just describe the moves.
Here is the honest state of it today:
- A person grades it, now. In a cohort or concierge run, a reviewer scores your support system against the master rubric above and tells you what is weak and what to revise. The rubric and the CX-lead bar are the product, and they are hand-gradeable today — that is the deliberate first step, not a placeholder.
- AI-assisted grading on the same rubric is next, not a claim we make yet. When it ships, it grades against this exact rubric — the bar does not move, the reviewer does.
- It is not an accreditation-body badge. Its weight is the work behind it: a complete, coherent support system a reviewer judged against a professional standard. That is a real thing to point at — and a more useful one than a quiz score.
No automated verification or paywall is wired behind this today that the line above does not describe. The certificate means what the work means.
What’s next
After the certificate, keep two things beside you as you run the system: the Support operating guide — the data, escalation, and sign-off layer that belongs underneath a team putting real customer replies through Claude — and the foundation toolkit from Module 1, the support-voice.md and support-policy.md files that every later piece inherits. When a new agent joins or a policy changes, update those two files and the whole system updates with them.
A manager dashboard — a team view of who is certified, queue coverage, and QA trend by agent — is in the roadmap and is not a claim we make today. When it ships, it lands directly into the operating system you built here.
The next capability track is your call. If your team also handles billing disputes, the Finance & Ops track covers the escalation and refund mechanics from the finance side. If you manage the agents who run the system, the People & HR track covers the hiring, onboarding, and review loop that keeps the team sharp.
Build the system, clear the rubric, and the certificate is yours — because the work is.