Function cockpit
Customer Service
Agents resolve the routine. Humans hold the difficult conversations.
Phase: Phase 3 — Functional orchestration
Resolve common issues end to end across knowledge, billing, inventory and scheduling systems, while routing vulnerability, ambiguity and emotion to humans who can genuinely help.
Trust score
0
Touchless
0%
Human review
0%
Override rate
0%
Capacity released
40-60% autonomous resolution initially, progressing only when repeat contact and outcomes stay strong
Accountable human
Chief Customer Officer
CoS agent: Customer Service Chief of Staff
Eleven decomposed sub-functions, each with its own estate rating
Sub-functions
Customer Service, decomposed
Nobody automates a function. They automate the sub-functions inside it, at different speeds, with different ceilings. Each one carries its own orchestrator, its own human owner, and its own honest limit.
Autonomy ceiling range
72%→94%
The binding constraint here is Quality & Coaching — capped at 72%.
Scoring at 100 percent coverage is a genuine agentic unlock. Coaching a named person on their performance is delivered by their manager, not a model.
Intake Orchestrator · human owner Director, Service Intake
estate rating
Owns contact work end to end inside Customer Service. Runs 3 lines across 16 stations, drains the repeatable share into the O2C tower, and holds 0 human gates.
The highest ceiling on the platform. Classification, entitlement and routing are fully machine-decidable and every misroute is self-correcting within one hop.
Units / wk
4,225
Agents
10
Lines
3
Cost / unit
$6.72
was $26
Tier-1 Resolution Orchestrator · human owner Director, Tier-1 Service
estate rating
Owns case work end to end inside Customer Service. Runs 3 lines across 18 stations, drains the repeatable share into the O2C tower, and holds 2 human gates.
Agents resolve the long tail end to end inside a value envelope. Credits above the envelope, and any account closure, hold for a human.
Units / wk
3,499
Agents
10
Lines
3
Cost / unit
$15
was $44
Tier-2 Technical Orchestrator · human owner Director, Technical Support
estate rating
Owns escalated case work end to end inside Customer Service. Runs 3 lines across 18 stations, drains the repeatable share into no shared tower, and holds 1 human gates.
Diagnosis and workarounds are agentic. Declaring a product defect commits engineering capacity and needs a human confirming the call.
Units / wk
7,356
Agents
10
Lines
3
Cost / unit
$8.57
was $25
Case Management Orchestrator · human owner Director, Service Operations
estate rating
Owns case work end to end inside Customer Service. Runs 3 lines across 17 stations, drains the repeatable share into the O2C tower, and holds 1 human gates.
Lifecycle mechanics, SLA clocks and dedupe are pure state management. Only genuinely stuck cases surface to a human.
Units / wk
15,863
Agents
10
Lines
3
Cost / unit
$4.01
was $16
Knowledge Orchestrator · human owner Head of Knowledge Management
estate rating
Owns article work end to end inside Customer Service. Runs 3 lines across 17 stations, drains the repeatable share into the AI-OPS tower, and holds 3 human gates.
Drafting and localization are agentic. Publishing an answer customers will act on requires a human accuracy sign-off — wrong knowledge scales instantly.
Units / wk
8,585
Agents
10
Lines
3
Cost / unit
$13
was $43
Dispatch Orchestrator · human owner Director, Field Service
estate rating
Owns field job work end to end inside Customer Service. Runs 3 lines across 19 stations, drains the repeatable share into the M2D tower, and holds 1 human gates.
Matching, routing and rescheduling are optimization problems agents solve better than dispatchers. Safety-critical jobs still route to a named engineer.
Units / wk
14,926
Agents
10
Lines
3
Cost / unit
$18
was $50
Warranty & Returns Orchestrator · human owner Director, Warranty & Returns
estate rating
Owns claim work end to end inside Customer Service. Runs 3 lines across 18 stations, drains the repeatable share into the O2C tower, and holds 1 human gates.
Warranty validation and return authorization are rule-driven with strong fraud signals. High-value or repeat-pattern claims route to a human reviewer.
Units / wk
5,298
Agents
10
Lines
3
Cost / unit
$6.86
was $22
Voice of Customer Orchestrator · human owner Head of Customer Experience
estate rating
Owns feedback theme work end to end inside Customer Service. Runs 3 lines across 18 stations, drains the repeatable share into no shared tower, and holds 2 human gates.
Ingestion and theme extraction are agentic at full volume. Turning a theme into an enterprise commitment is a leadership decision.
Units / wk
3,997
Agents
10
Lines
3
Cost / unit
$13
was $45
Quality Orchestrator · human owner Director, Service Quality
estate rating
Owns quality review work end to end inside Customer Service. Runs 3 lines across 16 stations, drains the repeatable share into no shared tower, and holds 3 human gates.
Scoring at 100 percent coverage is a genuine agentic unlock. Coaching a named person on their performance is delivered by their manager, not a model.
Units / wk
14,523
Agents
10
Lines
3
Cost / unit
$8.65
was $22
Workforce Management Orchestrator · human owner Director, Workforce Management
estate rating
Owns roster work end to end inside Customer Service. Runs 3 lines across 16 stations, drains the repeatable share into no shared tower, and holds 2 human gates.
Forecasting and intraday rebalancing are agentic. Publishing a roster that governs peoples working lives holds a human gate and a works-council check.
Units / wk
6,810
Agents
10
Lines
3
Cost / unit
$3.33
was $11
Service Analytics Orchestrator · human owner Head of Service Analytics
estate rating
Owns analysis work end to end inside Customer Service. Runs 3 lines across 18 stations, drains the repeatable share into the DATA-ANALYTICS tower, and holds 3 human gates.
Assembly and driver analysis are agentic end to end. The narrative that goes to the executive committee carries a named human author.
Units / wk
119
Agents
10
Lines
3
Cost / unit
$738
was $2K
How to read an estate rating
The rating is not a maturity badge. It scores seven control dimensions — control coverage, evidence quality, override discipline, data integrity, recovery readiness, cost transparency, and judgment readiness. Judgment readiness runs deliberately against autonomy: a sub-function that has pushed automation hardest usually scores worst on whether a human could still take the decision back.