| name | service-blueprint |
| description | Create a service blueprint -- mapping both the customer-facing journey and the backstage organisational processes that support it. Extends journey mapping with employee actions, support systems, and physical evidence. Use when analysing end-to-end service delivery or identifying backstage bottlenecks. |
| argument-hint | [service or customer journey to blueprint] |
| user-invocable | true |
| allowed-tools | Read, Write, Edit, Bash, Glob, Grep |
Service Blueprint
Create a service blueprint for $ARGUMENTS. A service blueprint extends a journey map by adding everything that happens behind the scenes — employee actions, support systems, and handoffs that the customer never sees but that determine whether the experience works.
Reference: NNGroup Service Blueprints
Step 1: Define scope (mandatory)
### Blueprint scope
| Element | Detail |
|---|---|
| **Service** | [specific service being blueprinted] |
| **Journey** | [which customer journey — start to end] |
| **Start point** | [concrete trigger — e.g. "customer submits support ticket"] |
| **End point** | [concrete outcome — e.g. "issue resolved and customer confirms"] |
| **Customer type** | [who is the customer in this journey] |
| **Success metric** | [how you measure this service works — resolution time, NPS, completion rate] |
Rules for scope:
- One service, one blueprint. "Our entire product experience" is not a scope — "new customer onboarding from signup to first value" is
- Start and end must be concrete events with a clear trigger and outcome
- If a journey map exists from
/ux-researcher:journey-map, use it as the frontstage foundation
Output: Scope definition table.
Step 2: Map customer actions (mandatory)
Map every action the customer takes from start to end:
### Customer actions (top lane)
| Step | Customer action | Touchpoint | Channel |
|---|---|---|---|
| 1 | [what the customer does] | [what they interact with] | [web, app, email, phone, in-person] |
| 2 | [next action] | [touchpoint] | [channel] |
| ... | ... | ... | ... |
Rules for customer actions:
- These are ONLY things the customer does or sees. If the customer doesn't do it or see it, it belongs in a later lane
- Include wait times as explicit steps — "Customer waits for response (avg 4h)" is a real step
- Every touchpoint must specify the channel — "email" and "in-app notification" serve different purposes
Output: Complete customer action lane.
Step 3: Map frontstage employee actions (mandatory)
Map what employees do that the customer CAN see:
### Frontstage employee actions
| Step | Corresponding customer action | Employee action | Role | Touchpoint |
|---|---|---|---|---|
| 1 | [customer step] | [what the employee does visibly] | [role — support agent, CS manager, sales] | [email, call, chat, in-person] |
| 2 | [customer step] | [employee action] | [role] | [touchpoint] |
Rules for frontstage:
- Frontstage means the customer is aware of it. The support agent replying to a ticket is frontstage. The support agent looking up the customer's history is backstage
- Every customer action should have a corresponding frontstage response (even if automated). If a customer action has no response, that's a gap
- Specify the role, not the person — roles are replaceable, people are not
Output: Frontstage employee action lane aligned to customer actions.
Step 4: Draw the line of visibility (mandatory)
### Line of visibility
───────────────────── LINE OF VISIBILITY ─────────────────────
Everything above: the customer sees it
Everything below: the customer does NOT see it
**Visibility audit:**
| Customer action | What customer sees | What customer does NOT see |
|---|---|---|
| [action] | [frontstage response] | [backstage work that makes it possible] |
| [action] | [visible response] | [hidden process] |
Rules for the line of visibility:
- The line of visibility is the most important concept in the blueprint. It separates the experience (above) from the operations (below)
- Anything that crosses the line (backstage work that becomes visible) is a moment of truth — handle with care
- If the customer can see internal process leaking through (e.g. "your ticket has been escalated to tier 2"), that's a visibility breach. Decide if it's intentional or accidental
Output: Line of visibility with visibility audit table.
Step 5: Map backstage employee actions (mandatory)
Map what employees do that the customer CANNOT see:
### Backstage employee actions
| Step | Triggered by | Employee action | Role | System/tool used | Duration |
|---|---|---|---|---|---|
| 1 | [customer or frontstage action] | [what happens behind the scenes] | [role] | [internal tool] | [how long] |
| 2 | [trigger] | [backstage action] | [role] | [tool] | [duration] |
Rules for backstage:
- Every backstage action must be triggered by something — a customer action, a frontstage action, or another backstage action. No orphaned steps
- Include handoffs between people/teams as explicit steps — handoffs are where services break
- Duration matters — a "quick review" that takes 3 days is a bottleneck, not a quick review
Output: Backstage employee action lane with triggers and durations.
Step 6: Map support processes (mandatory)
Map the internal systems, tools, and infrastructure that enable the service:
### Support processes (bottom lane)
| Backstage action | Support system | Type | Owner | SLA/availability |
|---|---|---|---|---|
| [backstage step] | [system that enables it] | [CRM, database, API, queue, manual process] | [team] | [uptime, response time] |
| [backstage step] | [system] | [type] | [owner] | [SLA] |
Rules for support processes:
- If a backstage action relies on a system, that system belongs here. If the system goes down, the service fails
- "Manual process" is a valid support process — and often the most fragile. Flag it
- Identify single points of failure — systems or people where there is no backup
Output: Support process lane linked to backstage actions.
Step 7: Identify failure points (mandatory)
### Failure points
| # | Location | Failure mode | Impact on customer | Frequency | Root cause | Current mitigation |
|---|---|---|---|---|---|---|
| F1 | [lane and step] | [what goes wrong] | [what the customer experiences] | [daily/weekly/rare] | [why it fails] | [what's in place today — or "none"] |
| F2 | [location] | [failure] | [customer impact] | [frequency] | [cause] | [mitigation] |
Focus areas for failure points:
- Handoffs between people — information lost, delays, miscommunication
- Handoffs between systems — data sync failures, API timeouts, format mismatches
- Line of visibility crossings — internal errors leaking to customers
- Manual processes — human error, inconsistency, capacity bottlenecks
- Wait times — any step where the customer waits more than expected
Output: Failure point table with root causes and current mitigations.
Step 8: Recommend improvements (mandatory)
### Recommendations (prioritised by customer impact)
| Priority | Failure point | Recommendation | Impact | Effort | Owner |
|---|---|---|---|---|---|
| 1 | [F#] | [specific improvement] | [how it improves the customer experience] | [S/M/L] | [team] |
| 2 | [F#] | [improvement] | [impact] | [effort] | [team] |
| 3 | [F#] | [improvement] | [impact] | [effort] | [team] |
Rules for recommendations:
- Prioritise by customer impact, not ease of implementation
- "Automate X" is only a recommendation if you specify what to automate and how
- Quick wins (high impact, low effort) go first, but don't ignore systemic issues
Output: Prioritised recommendation table.
Rules
- The line of visibility is the key insight. It separates what the customer sees from what they don't. Everything interesting happens at or near this line.
- Every customer action must have a corresponding backstage process. If there's no backstage support, who's handling it? Either someone is and you haven't mapped them, or no one is and that's a failure point.
- Failure points at handoffs are the most common. When work passes from one person to another, or one system to another, information is lost and delays accumulate. Look at handoffs first.
- Wait times are steps, not gaps. A customer waiting 48 hours for a response is experiencing the service. Map it, measure it, improve it.
- Manual processes are fragile. Every manual step is a reliability risk. Don't assume manual means simple — it often means inconsistent.
- Don't blueprint what doesn't exist. Map the current state first. Aspirational blueprints are roadmaps, not blueprints.
Output Format
# Service Blueprint: [service name]
## Scope
[From Step 1]
## Customer Actions
[Lane from Step 2]
## Frontstage Employee Actions
[Lane from Step 3]
───────────────────── LINE OF VISIBILITY ─────────────────────
## Backstage Employee Actions
[Lane from Step 5]
## Support Processes
[Lane from Step 6]
## Failure Points
[Table from Step 7]
## Recommendations
[Prioritised table from Step 8]
---
Service: [name]
Journey: [start → end]
Failure points identified: [N]
Last updated: [date]
Related Skills
/ux-researcher:journey-map — the customer-facing layer of the blueprint. If a journey map exists, use it as the foundation and add the backstage lanes.
/ux-researcher:usability-review — for deep-diving into specific frontstage touchpoints that the blueprint identifies as high-friction.
Use the service blueprint template (templates/service-blueprint.md) for output structure.