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
Strategy, policy, observe, workflow, and AI GBS side by side
Five layers
What Customer Service actually runs
A function is not one thing. It sets direction, writes the rules it is bound by, watches the world and itself, does the work on lines, and pushes repeatable transactions into a shared spine. Agents work in all five — not just the last two.
Strategy
14 agents
Agents that take positions, argue them, and get held to the outcome.
14 positions on the record
3 ratified, 1 in challenge
Policy
15 agents
The written rules every other agent is bound by, with version and owner.
31 binding rules
25 active, 6 superseded or retired
Observe
19 agents
Continuous sensing of the outside world and of the agents themselves.
19 watchers running
They watch the outside world and the agents themselves
Workflow
45 agents
Doing the work: lines, stations, and live units moving through them.
33 lines, 161 live units
14 held for a person right now
AI GBS
17 agents
Repeatable transaction processing run once for the whole enterprise.
4 shared towers
46 stations on this function run inside AI GBS
Strategy layer — positions on the record
Every position names its author, the agent paid to argue against it, and the conditions that would break it.
Intake & Routing Operating Position
SupersededHold the ceiling at A4 for intake & routing. The highest ceiling on the platform. Classification, entitlement and routing are fully machine-decidable and every misroute is self-correcting within one hop. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.
Tier-1 Resolution Operating Position
RatifiedHold the ceiling at A4 for tier-1 resolution. Agents resolve the long tail end to end inside a value envelope. Credits above the envelope, and any account closure, hold for a human. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.
Tier-2 Technical Operating Position
RejectedHold the ceiling at A3 for tier-2 technical. Diagnosis and workarounds are agentic. Declaring a product defect commits engineering capacity and needs a human confirming the call. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.
Case Management Operating Position
RejectedHold the ceiling at A4 for case management. Lifecycle mechanics, SLA clocks and dedupe are pure state management. Only genuinely stuck cases surface to a human. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.
Knowledge Doctrine Review
In challengeHold the ceiling at A3 for knowledge management. Drafting and localization are agentic. Publishing an answer customers will act on requires a human accuracy sign-off — wrong knowledge scales instantly. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.
Knowledge Management Operating Position
SupersededKeep the repeatable article volume in the spine and pull the exception tail back into the function. The spine is cheaper per unit; the tail is where the context that prevents rework actually sits.
Policy layer — what binds the agents
Versioned rules with an owning agent, a human approver, and the agents they constrain.
Intake & Routing Decision Envelope
vv2.1Agents in this lane may act without a human when the contact sits inside the stated value, risk and confidence envelope. Outside it, the unit holds at a gate with a named approver and a running clock.
Intake & Routing Evidence Standard
vv1.0Every autonomous decision writes inputs, the rule version applied, the model and prompt version, the output and a reversal path. An action with no evidence record is treated as a control failure, not a fast decision.
Intake & Routing Escalation Rule
vv2.5Escalation is mandatory when confidence falls below the floor, when two options sit inside the confidence band, or when a break condition on a ratified position fires. Director, Service Intake owns the response clock.
Intake & Routing Autonomy Ceiling
vv4.5The ceiling for this lane is 94 on the platform scale. The highest ceiling on the platform. Classification, entitlement and routing are fully machine-decidable and every misroute is self-correcting within one hop. The ceiling is a governance decision, not a technical limit, and only Director, Service Intake can propose moving it.
Tier-1 Resolution Decision Envelope
vv5.8Agents in this lane may act without a human when the case sits inside the stated value, risk and confidence envelope. Outside it, the unit holds at a gate with a named approver and a running clock.
Tier-1 Resolution Evidence Standard
vv3.1Every autonomous decision writes inputs, the rule version applied, the model and prompt version, the output and a reversal path. An action with no evidence record is treated as a control failure, not a fast decision.
Workflow layer — the lines doing the work
Each line is a workflow. Each station is a stage. Each unit is a live piece of work.
| Line | Sub-function | Kind | Units / wk | Touchless | Timetable | Actual | Stations |
|---|---|---|---|---|---|---|---|
| CS1A Contact to Routed Case | Intake & Routing | transactional | 2,640 | 95% | 1.6h | 1.9h | 6 |
| CS1B Self-Service Deflection | Intake & Routing | transactional | 463 | 92% | 16.8h | 13.0h | 5 |
| CS1C Surge Absorption | Intake & Routing | transactional | 1,122 | 93% | 18.0h | 21.0h | 5 |
| CS2A Query to Resolution | Tier-1 Resolution | transactional | 916 | 91% | 13.0h | 14.4h | 6 |
| CS2B Billing Dispute | Tier-1 Resolution | transactional | 1,787 | 89% | 3.5h | 3.1h | 6 |
| CS2C Goodwill Credit | Tier-1 Resolution | transactional | 796 | 89% | 15.1h | 17.1h | 6 |
| CS3A Escalated Case to Fix | Tier-2 Technical | transactional | 5,568 | 70% | 8.6h | 7.3h | 7 |
| CS3B Defect Handoff | Tier-2 Technical | transactional | 1,700 | 80% | 16.1h | 19.6h | 6 |
| CS3C Recurring Issue Analysis | Tier-2 Technical | analytical | 88 | 53% | 129.0h | 133.6h | 5 |
| CS4A Case Lifecycle | Case Management | transactional | 5,193 | 86% | 24.1h | 27.2h | 7 |
| CS4B Backlog Sweep | Case Management | transactional | 2,856 | 90% | 21.8h | 18.7h | 5 |
| CS4C Reopen Handling | Case Management | transactional | 7,814 | 96% | 1.5h | 1.2h | 5 |
| CS5A Gap to Published Article | Knowledge Management | transactional | 1,401 | 84% | 2.6h | 3.2h | 6 |
| CS5B Freshness Cycle | Knowledge Management | transactional | 7,182 | 74% | 9.5h | 9.8h | 5 |
| CS5C Knowledge Doctrine Review | Knowledge Management | judgment | 2 | 49% | 723.7h | 655.4h | 6 |
| CS6A Job to Completion | Field Service Dispatch | transactional | 5,851 | 84% | 14.9h | 16.8h | 9 |
| CS6B Same-day Re-optimization | Field Service Dispatch | transactional | 5,732 | 84% | 2.9h | 3.2h | 5 |
| CS6C Parts Escalation | Field Service Dispatch | transactional | 3,343 | 83% | 10.0h | 10.2h | 5 |
| CS7A Claim to Resolution | Warranty & Returns | transactional | 2,755 | 91% | 5.9h | 4.5h | 7 |
| CS7B Supplier Recovery | Warranty & Returns | transactional | 2,538 | 84% | 25.7h | 33.4h | 5 |
| CS7C Repair vs Replace Position | Warranty & Returns | judgment | 5 | 53% | 725.7h | 799.8h | 6 |
| CS8A Feedback to Insight | Voice of Customer | analytical | 12 | 57% | 142.0h | 154.5h | 6 |
| CS8B Detractor Closed Loop | Voice of Customer | transactional | 3,980 | 82% | 14.6h | 18.0h | 6 |
| CS8C Experience Strategy Position | Voice of Customer | judgment | 5 | 40% | 858.7h | 694.8h | 6 |
| CS9A Sample to Coaching Action | Quality & Coaching | transactional | 8,078 | 72% | 15.2h | 11.8h | 6 |
| CS9B Compliance Screening | Quality & Coaching | transactional | 6,388 | 68% | 24.2h | 30.4h | 5 |
| CS9C Calibration Cycle | Quality & Coaching | analytical | 57 | 57% | 49.9h | 52.5h | 5 |
| CS10A Forecast to Schedule | Workforce Management | analytical | 54 | 70% | 169.7h | 202.7h | 6 |
| CS10B Intraday Rebalance | Workforce Management | transactional | 6,682 | 84% | 1.4h | 1.4h | 5 |
| CS10C Outsourcer Allocation | Workforce Management | analytical | 74 | 71% | 151.9h | 188.7h | 5 |
| CS11A Period Service Review | Service Analytics | analytical | 53 | 64% | 153.4h | 135.7h | 6 |
| CS11B Contact Driver Reduction | Service Analytics | analytical | 61 | 58% | 56.1h | 44.4h | 6 |
| CS11C Cost-to-Serve Position | Service Analytics | judgment | 5 | 48% | 702.0h | 707.7h | 6 |
Observe layer — the watchers
Sensing agents and the agents that supervise other agents.
Intake Orchestrator
Owns the intake & routing lane end to end. Sequences the other agents, holds the timetable, and decides what surfaces to a human.
Intake & Routing Challenger Agent
Adversarial reviewer for intake & routing. Argues the opposite case on every position and flags where the evidence does not carry the claim.
Tier-1 Resolution Orchestrator
Owns the tier-1 resolution lane end to end. Sequences the other agents, holds the timetable, and decides what surfaces to a human.
Sentiment Monitor Agent
Watches tier-1 resolution continuously. Detects drift, quantifies it, and routes what matters without waiting for a reporting cycle.
Tier-1 Resolution Challenger Agent
Adversarial reviewer for tier-1 resolution. Argues the opposite case on every position and flags where the evidence does not carry the claim.
Tier-2 Technical Orchestrator
Owns the tier-2 technical lane end to end. Sequences the other agents, holds the timetable, and decides what surfaces to a human.
Tier-2 Technical Challenger Agent
Adversarial reviewer for tier-2 technical. Argues the opposite case on every position and flags where the evidence does not carry the claim.
Case Management Orchestrator
Owns the case management lane end to end. Sequences the other agents, holds the timetable, and decides what surfaces to a human.
AI GBS layer — the shared spine
Repeatable transaction processing does not belong to this function. It runs once for the enterprise.
Intake & Routing · Tier-1 Resolution · Case Management · Warranty & Returns
Knowledge Management
Field Service Dispatch
Service Analytics
Stations running inside AI GBS
46 of 191 stations(24% of the function's stations)
Read this page as the honest answer to “what would an 80% agentic Customer Service look like?” The strategy and policy layers are where the argument happens. The workflow and AI GBS layers are where the volume goes. The observe layer is the only reason the other four can be trusted.
Sub-function breakdown