LensReading which lens this session carries.
AI GBS & Global Business Services/Cost-to-Serve Management

Sub-function

Cost-to-Serve Management

Owns cost line work end to end inside AI GBS & Global Business Services. Runs 3 lines across 19 stations, drains the repeatable share into the R2R tower, and holds 2 human gates.

orchestrator Cost-to-Serve Orchestratorhuman owner Head of Service Economics routes into r2r

Estate rating

BBB

72 / 100

autonomy 84%cap 85%

The tower that keeps the platform honest. Computation is agentic; the published unit economics carry a named human author because they are used to make investment decisions.

Agents

11

Lines

3

Units / wk

8,133

Timetable vs actual

20.0h → 19.4h

Cost / unit

$2.85

was $11

Run-rate saving

$3.3M

annualized at this volume

The 11 agents running this sub-function

Grouped by what they are for, not by where they sit. The agent org is process-shaped.

OrchestratorOwns the queue and arbitrates between desks.

Cost-to-Serve Orchestrator

Owns the cost-to-serve management lane end to end. Sequences the other agents, holds the timetable, and decides what surfaces to a human.

ObserveA355 / 24h88%7% ovr$5.55/task

ceiling A3 · escalates to Escalates to Head of Service Economics when the unit falls outside the decision envelope or confidence drops below the floor.

AI GBSRuns the shared transaction spine.

Unit Cost Agent

Runs the repeatable portion of cost-to-serve management inside the shared transaction spine at spine cost and spine controls.

AI GBSA32,213 / 24h94%8% ovr$0.10/task

ceiling A3 · escalates to Escalates to Head of Service Economics when the unit falls outside the decision envelope or confidence drops below the floor.

Chargeback Agent

Runs the repeatable portion of cost-to-serve management inside the shared transaction spine at spine cost and spine controls.

AI GBSA32,342 / 24h88%1% ovr$0.19/task

ceiling A3 · escalates to Escalates to Head of Service Economics when the unit falls outside the decision envelope or confidence drops below the floor.

Cost Narrative Agent

Runs the repeatable portion of cost-to-serve management inside the shared transaction spine at spine cost and spine controls.

AI GBSA32,385 / 24h93%6% ovr$0.13/task

ceiling A3 · escalates to Escalates to Head of Service Economics when the unit falls outside the decision envelope or confidence drops below the floor.

TaskExecutes stations on a line.

Inference Cost Agent

Executes the cost-to-serve management step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA3915 / 24h93%5% ovr$0.44/task

ceiling A3 · escalates to Escalates to Head of Service Economics when the unit falls outside the decision envelope or confidence drops below the floor.

Volume Attribution Agent

Executes the cost-to-serve management step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA31,828 / 24h97%8% ovr$0.26/task

ceiling A3 · escalates to Escalates to Head of Service Economics when the unit falls outside the decision envelope or confidence drops below the floor.

Benchmark Agent

Executes the cost-to-serve management step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA33,442 / 24h89%4% ovr$0.45/task

ceiling A3 · escalates to Escalates to Head of Service Economics when the unit falls outside the decision envelope or confidence drops below the floor.

StrategyForms and defends a position.

Margin of Automation Agent

Authors positions on cost-to-serve management. Frames the question, models the options, and writes a recommendation a human can ratify or reject.

StrategyA271 / 24h95%11% ovr$5.29/task

ceiling A3 · escalates to Escalates when two options sit inside the confidence band and the evidence cannot separate them.

Economics Doctrine Agent

Authors positions on cost-to-serve management. Frames the question, models the options, and writes a recommendation a human can ratify or reject.

StrategyA2116 / 24h94%11% ovr$9.02/task

ceiling A3 · escalates to Escalates when two options sit inside the confidence band and the evidence cannot separate them.

PolicyAuthors and versions the binding rules.

Chargeback Policy Agent

Holds the rules that bind every other agent in cost-to-serve management. Versions them, tests them against live decisions, and blocks work that breaches them.

PolicyA295 / 24h91%4% ovr$3.49/task

ceiling A3 · escalates to Escalates on any decision that would breach a live policy version, with the clause cited.

ChallengerPaid to disagree before a human has to.

Cost-to-Serve Management Challenger Agent

Adversarial reviewer for cost-to-serve management. Argues the opposite case on every position and flags where the evidence does not carry the claim.

StrategyA2208 / 24h96%15% ovr$8.43/task

ceiling A3 · escalates to Escalates when a ratified position is still running against a break condition it flagged.

Estate rating breakdown

Seven control dimensions. Judgment readiness runs against autonomy on purpose.

Control Coverage72
Evidence Quality73
Override Discipline79
Data Integrity85
Recovery Readiness66
Cost Transparency75
Judgment Readiness57

People

What the humans stopped doing, and what they do now.

headcount on this work60.233.2 FTE

Redeployed into exception judgment and supplier relationship work.

inference cost per unit$0.66

Lines running in this sub-function

Each line is a workflow. Click a line to walk it station by station.

6 clear5 evidenced0 held
GB11APeriod Cost-to-ServeAnalytical · 7 stations

Collect volumes

2 q

Collect model spend

35 q

Compute unit cost

13 q

Compare baseline

29 q

Draft narrative

1 live

Review

1 live

Publish

27 q

GB11BChargeback RunTransactional · 5 stations

Allocate

1 live

Validate

1 live

Dispute window

3 live

Resolve

1 live

Post

1 live

GB11CAutomation Economics PositionJudgment · 7 stations

Frame question

27 q

Gather evidence

28 q

Model options

1 live

Draft position

39 q

Challenge

31 q

Revise

1 live

Ratify

4 q

Strategy positions this sub-function holds

An agent that only executes is a robot. These are the calls it made, the argument against each one, and what would prove it wrong.

GB11-POS-100

Automation Economics Position

Rejectedv3

Which parts of cost-to-serve management should deliberately stay human, and how do we defend that against a cost argument?

Hold the ceiling at A3 for cost-to-serve management. The tower that keeps the platform honest. Computation is agentic; the published unit economics carry a named human author because they are used to make investment decisions. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.

confidence

61%

The challenge — Cost-to-Serve Management Challenger Agent

Cost-to-Serve Management Challenger Agent argues the recommendation leans on 3 quarters of data from a period with no volume shock. The confidence band overlaps the alternative, and the position does not say what it would take to be wrong. Recorded as a dissent, not a block.

Break conditions

  • ·Override rate rises above 14 percent for two consecutive months
  • ·Cost per cost line stops falling while volume keeps rising
  • ·A control failure in this lane reaches a customer or a regulator

Evidence gaps

  • ·No independent test of the control evidence behind the touchless rate
Alternative consideredCostRiskVerdict
Hold the current position$318k run rateKnown and priced. Cedes ground if comparators move faster.live
Raise the ceiling one level now$821k to build controlsOverride rate is 13 percent; raising the ceiling before that settles imports the error into production.rejected on evidence
Move the exception tail to the spine$1089k transitionLoses local context. Rework risk on the cases that are hardest to recover.under review
authored by Margin of Automation Agentratifier Head of Service Economics3 stated assumptions
GB11-POS-101

Cost-to-Serve Management Operating Position

In challengev1

Where should the autonomy ceiling for cost-to-serve management sit over the next four quarters, and what evidence would justify moving it?

Keep the repeatable cost line 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.

confidence

86%

The challenge — Cost-to-Serve Management Challenger Agent

Cost-to-Serve Management Challenger Agent argues the recommendation leans on 3 quarters of data from a period with no volume shock. The confidence band overlaps the alternative, and the position does not say what it would take to be wrong. Recorded as a dissent, not a block.

Break conditions

  • ·Override rate rises above 15 percent for two consecutive months
  • ·Cost per cost line stops falling while volume keeps rising
  • ·A comparator publishes a materially lower unit cost on a like-for-like basis

Evidence gaps

  • ·No external comparator on a like-for-like unit definition
  • ·Exception cases under $46k are sampled, not fully measured
Alternative consideredCostRiskVerdict
Hold the current position$829k run rateKnown and priced. Cedes ground if comparators move faster.live
Raise the ceiling one level now$922k to build controlsOverride rate is 10 percent; raising the ceiling before that settles imports the error into production.rejected on evidence
Move the exception tail to the spine$95k transitionLoses local context. Rework risk on the cases that are hardest to recover.under review
authored by Margin of Automation Agentratifier Head of Service Economics3 stated assumptions

Policy register

The rules the agents above are bound by. Version, owner, approver, and the agents each rule constrains.

RefRuleScopeOwner agentApproved byVersionStatus
GB11-POL-200

Cost-to-Serve Management Decision Envelope

Agents in this lane may act without a human when the cost line 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.

binds 6 agents · EU · UK

envelopeChargeback Policy AgentHead of Service Economicsv1.9Active
GB11-POL-201

Cost-to-Serve Management Evidence Standard

Every 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.

binds 6 agents · APAC

evidenceChargeback Policy AgentHead of Service Economicsv2.6Active
GB11-POL-202

Cost-to-Serve Management Escalation Rule

Escalation 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. Head of Service Economics owns the response clock.

binds 6 agents · APAC

escalationChargeback Policy AgentHead of Service Economicsv1.9Active