TemplateClone Runbook
The operating manual for one production Clone, covering what it does, how it is configured, routine checks, common failures, change control and who to call.
#What this is for
When the Clone team steps back, your people need to run every production Clone confidently: know what it is supposed to do, spot when it is not, fix the common problems and change it safely. A runbook per Clone makes that possible. It is written for the org admin or CoE builder who was not in the room when the Clone was designed, and it is the core of the handoff documentation at Month 12.
#When to use it
Start a runbook at each Clone's go-live; the go-live checklist already captures most of section 2. Complete and verify all runbooks in Months 10–12. Update a runbook whenever the Clone's skills, connections, schedules or autonomy level change.
#Who's involved
- Implementation Engineer writes the first version.
- Org admin and CoE builders become the owners and verify it by running it.
- Business owner confirms purpose, limits and escalation.
- Champion adds the common user questions.
#How to fill it in
- Copy the template for each production Clone.
- Fill in sections 1–3 from the Clone Design Spec, Governance Charter and Integration Architecture. Link rather than duplicate where you can, but keep the essentials here.
- Write routine operations (section 4) as steps someone can follow without help.
- Build the troubleshooting table (section 5) from real issues seen in pilot and hypercare.
- Record change control and rollback (section 6).
- Verify: a CoE member runs one routine check and one troubleshooting step using only the runbook. Fix whatever they had to ask about.
#The template
#1. Clone profile
| Field | Entry |
|---|
| Clone name | Billing Clone |
| Department | Finance & Accounting |
| Role / purpose | Researches billing disputes and issues credit memos within limits |
| Business owner | |
| Escalation owner | AR Manager |
| Champion | |
| Technical owner (org admin / CoE) | |
| Created from | Marketplace template + 1 custom skill |
| Autonomy level(s) by workflow | Credit memos within limits: 4; write-offs: 2 |
| Channels | Slack #ar-billing, email, Commander handoffs |
| Governance Charter entry | Link |
| Design spec version | v2.1 |
#2. Configuration
Skills
| Skill | Source | Version (pinned) | Purpose | Test pack |
|---|
| ERP AR Posting | Custom (SkillForge) | v1.1 | Create credit memos | 12 scenarios |
| Helpdesk | Marketplace | v2.3 | Read dispute tickets | 6 scenarios |
Connections
| System | Account type | Scope | Approved by | Review due |
|---|
| ERP | Service account | AR read; credit memo create; no payments | InfoSec lead | |
Knowledge
| Source | What it covers | Owner | Refresh |
|---|
| Credit policy v4 | Limits, reasons, approvals | Controller | On change |
Schedules and triggers
| Name | When | What it does |
|---|
| Morning dispute sweep | Weekdays 7:00 | Pulls new billing tickets and prepares research |
| Handoff from Support Clone | Event | Starts dispute workflow |
#3. Limits and boundaries
| Item | Entry |
|---|
| Acts on its own when | Credit within agreed threshold, existing customer, invoice under 120 days, one credit per invoice |
| Asks a person when | Anything outside a limit; customer mentions legal action |
| Never allowed | Change bank details; issue payments; write off balances |
| Data it must not use | Card numbers; employee data |
#4. Routine operations
| Task | Frequency | Steps | Healthy looks like | Owner |
|---|
| Check Clone Health | Daily (weekdays) | Open the Clone in Clone Team Studio; review health and failed tasks | No failed tasks older than 1 day | Org admin |
| Review exception digest | Daily | Read digest in #ar-billing; resolve or reassign | All exceptions actioned same day | AR Manager |
| Sample audit | Monthly | Pull 5% of within-limits actions; check against policy; log result | 98%+ correct | AR Manager + CoE |
| Power Rating and usage review | Monthly | Review trend; note drops; check credit usage | Stable or rising; usage in line with volume | CoE |
| Knowledge freshness | Quarterly or on policy change | Confirm sources are current; update | No stale policy | Business owner |
| Connection and access review | Quarterly | Confirm scopes and users match the plan | No drift | InfoSec + org admin |
#5. Troubleshooting
| Symptom | Likely cause | Check | Fix | Escalate to |
|---|
| Clone says it cannot reach the ERP | Expired or revoked connection | Connection status in settings | Reconnect with the approved service account | IT lead if account locked |
| Wrong cost center or reason codes | Stale knowledge | Compare to current credit policy | Update knowledge; re-run test scenarios | Business owner |
| Skill errors after a system upgrade | Upstream API change | Skill error message; recent system changes | Roll back to prior pinned version or pause workflow; raise with lead builder | Lead builder |
| Clone acted outside a limit | Limit misconfigured or new edge case | Audit log entry | Pause the workflow; apply demotion trigger; incident review | InfoSec lead + business owner |
| Scheduled task did not run | Schedule disabled or credits | Schedule settings; credit balance | Re-enable; review credit settings | Org admin |
| Users say answers are vague | Instructions or knowledge gap | Read recent conversations | Tighten instructions; add examples | Champion |
#6. Change control and rollback
- Every change goes through the CoE change review (skill version, connection, schedule, limits, instructions)
- Skill upgrades are tested against the test pack before production
- New connections and scope changes are approved by InfoSec
- Autonomy changes require an evidence pack and dual sign-off (InfoSec lead and business owner)
| Rollback action | How | Who can do it |
|---|
| Revert a skill | Re-pin previous version | Lead builder / org admin |
| Reduce autonomy | Restore approval rule; drop one level | Org admin, on request of InfoSec or business owner |
| Pause the Clone | Pause in Clone Team Studio; notify Champion and escalation owner | Org admin, InfoSec lead, business owner |
| Cut access | Revoke the connection | Org admin, IT lead |
Change log:
| Date | Change | Version | Approved by | Tested | Rollback to |
|---|
| | | | | |
| Need | Who | How |
|---|
| Business question about the Clone's work | Business owner | |
| User help | Champion | |
| Technical fix | Org admin / CoE | |
| Security concern or incident | InfoSec lead | |
| Platform support | Clone support | support.clone.org |
#8. Handoff verification
- A CoE member ran one routine check using only this runbook
- A CoE member resolved one troubleshooting scenario using only this runbook
- Rollback rehearsed in a test or low-risk window
- All links work and owners are current
#Good looks like
- Written for someone who was not there. If a step needs a verbal explanation, the step is incomplete.
- Troubleshooting comes from real incidents, not imagined ones.
- Limits and "never allowed" are repeated here, not only in the charter. Operators need them at hand.
- Rollback is specific and has been rehearsed.
- Owners are people with names, not teams.
- Common mistake: one runbook for "all Clones". Each production Clone needs its own; shared procedures can be linked.
#How Clone helps
The Enablement Clone drafts each runbook from the design spec, Governance Charter, Integration Architecture and hypercare log, and keeps it current when configuration changes. It also answers operators' how-to questions with the runbook as its source. The Project Clone tracks runbook completion and handoff verification per Clone in the portal, which feeds the Handoff Documentation deliverable.
#Sign-off
| Role | Name | Date | Decision |
|---|
| Business owner | | | Purpose and limits confirmed |
| CoE lead | | | Accepted for operation |
| Org admin | | | Handoff verified |
| Implementation Engineer | | | Handed over |