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
Autonomy, trust, and throughput at a glance
Health
Is this function holding
Four readings with a printed rule behind each, then the outcome measures the function is judged on. A reading without a rule is decoration.
Autonomy against target
72%
10 points short of the 82% target
Trust score
86
at or above the 85 floor for a rung promotion
Touchless rate
68%
32% of items still take a person somewhere in the run
Override rate
8%
people are reversing the agents often enough that the ceiling is doing real work
Autonomous resolution
+1668%
Cases closed without human handling
Repeat contact
-4.27.4%
Customers returning within seven days on the same issue
Vulnerable routing accuracy
+398%
Correctly transferred to a trained human specialist
Cost per resolution
-0.621.84$
Blended compute, tooling and human handling
Customer satisfaction
+0.34.5/5
Post-resolution survey across all channels
Operations
What the function is running
The roster and where it sits on the ladder, the split between what runs alone and what a person still signs, and a slice of the floor as it stands.
Where this roster sits on the autonomy ladder
A0 assisted through A4 autonomous. Moving an agent up a rung is a governance decision, not a config change.
Agents
10
Active now
9
Tasks / 24h
42,518
Mean success
93%
Work mix, as it stands
How the function's volume divides between the agents and the people.
Runs end to end without a person
68%
cleared inside the ceiling, no queue, no signature
A person reads it before it clears
24%
evidence posted, a named reviewer signs
A person reverses the agent
8%
the agent proposed, the human decided otherwise
Capacity released
40-60% autonomous resolution initially, progressing only when repeat contact and outcomes stay strong
What runs without a person
Committed to autonomy inside a defined ceiling.
- Identity-appropriate information and order status
- Standard refunds and remedies within published policy
- Rescheduling, rerouting and appointment changes
- Case documentation and disposition coding
- Proactive notification before a customer makes contact
What stays with people
Judgment, accountability, and anything a regulator would ask a human about.
- Hardship, bereavement and vulnerable customers
- Safety issues and legal threats
- Large refunds and fraud ambiguity
- Retention exceptions outside published remedies
- Reputational incidents and relationship recovery
On the floor right now
A slice of live work. The full board carries every item.
- CS-8841Complete
Hardship case — warm transfer completed
Recover · Specialist desk · 12m old · 1 customer
- CS-8842Awaiting human
Goodwill remedy above band — repeat failure
Resolve · Service manager · 1h 18m old · $1,240
- CS-8843Escalated
Defect cluster — firmware release 8.2
Understand · Engineering · 2h 48m old · 214 contacts
- CS-8844Autonomous
Proactive notification — delayed shipment cohort
Contact · Cust Svc CoS · 34m old · 1,840 customers
- CS-8845Autonomous
Billing correction batch — proration error
Resolve · Cust Svc CoS · 52m old · $28.4K
- CS-8846Drafted
Knowledge gap remediation — warranty topic
Verify · Content owner · 4h 6m old · 9 articles
Actions
What this function is asking a person to do
Ordered by size. Each figure is a live count from this function's own work and governance records, not a target.
2
Items awaiting a person
queued against a named human, clock running
2
Governance decisions pending
an agent stopped at a gate and asked
0
Items older than 48 hours
on the floor long enough to be a problem
0
Agents below 80 confidence
running, but not at a level that supports a promotion
Brakes available right now
What a named human can pull today to stop this function, without waiting for an engineer.
- Identity verification failure in a served interaction
- Harmful or unsafe advice detected in an agent response
- Repeated unresolved loop without human offer
- Sentiment escalation without a routing to a person
- Abnormal remedy or refund volume in a short window
Live observability
What has actually been decided, and where each agent stops
A dashboard that shows only outcomes hides the decisions that produced them. This is the governance record as written, and the ceiling every agent is held to.
Governance record
Most recent first. Each entry names the actor and the call.
Hardship case transferred within 40 seconds
ContainedVulnerability Agent detected bereavement language, stopped all automation on the case and warm-transferred the customer to a trained specialist with context preserved.
escalation · Vulnerability Agent · materiality high
Goodwill remedy above policy band pending
PendingResolution Agent assembled a $1,240 remedy for a repeated fulfillment failure. Held for service-manager decision with cost-to-serve and relationship history attached.
approval · Service manager · materiality medium
Systemic defect cluster raised to Engineering
PendingProactive Service Agent correlated 214 contacts to one firmware release and packaged the evidence for Engineering, Legal disclosure review and Communications.
escalation · Customer Service Chief of Staff · materiality high
Response path suspended on grounding failure
ContainedQuality Agent detected an answer without a valid source citation in a warranty topic. The response path was suspended and the knowledge gap routed to the content owner.
oversight · Quality Agent · materiality medium
Escalation ceilings
Past the line the agent stops and hands the decision to the named human with the evidence attached.
- A3
Customer Service Chief of Staff
Ceiling: Routing and coordination only — no remedy authority above policy
Then: Any systemic defect pattern escalates to Engineering and the CCO within the same operating cycle.
- A3
Intent Agent
Ceiling: Classification and routing only
Then: Distress, safety or legal language routes to a human immediately without further automation.
- A3
Knowledge Agent
Ceiling: Published content only; no improvised policy interpretation
Then: Content gaps are reported to the knowledge owner and the customer is offered a human.
- A3
Resolution Agent
Ceiling: Remedies up to $500 within published policy
Then: Any remedy above $500 or outside the published catalog requires a human decision.
- A3
Billing Agent
Ceiling: Corrections up to $500; larger adjustments route to Finance
Then: Suspected systemic billing error escalates to Finance and the service owner.
Is policy and strategy coming to fruition
Does the intent above this function reach the work inside it
Counted per sub-function, where each step only counts if the step before it did. A policy that never reaches a running workflow has not landed, however well it reads.
Chain from stance to transaction
3 of 11 sub-functions carry a ratified position, an active policy and a live workflow
Ratified position
3
of 11 sub-functions — a stance the leadership signed, not a draft
...and an active policy
3
a rule in force, with a version and an approver
...reaching a live workflow
3
the rule reaches something that actually runs
...and landing in a GBS tower
2
the run is executed on the shared transaction spine
8 of the 11 sub-functions are executing at volume without the full chain behind them. They run to a general standard rather than to a rule with a version and an approver, and the gap first appears at the position step.
What this page is, and what it is not
Every reading here is computed live from this function's own records, which makes it exact and makes it narrow. Volumes, unit costs, autonomy and trust are modeled: none has been reconciled against an enterprise resource planning system, a service management tool or a payroll register. Read the outcome measures as the shape of the argument, not as an audited result.
Gartner predicts that by 2029 agentic AI will autonomously resolve 80% of common customer-service issues and reduce operating costs by 30%. This is a dated prediction, not current performance. Gartner also predicts generative-AI cost per resolution could exceed $3 by 2030 and that regulatory changes could increase assisted-service volume by 30% by 2028. BCG reports up to 50% lower time to resolution in selected implementations.