LensReading which lens this session carries.
Engineering/Test & Assurance

Sub-function

Test & Assurance

Owns test run work end to end inside Engineering. Runs 3 lines across 15 stations, drains the repeatable share into the C2L tower, and holds 3 human gates.

orchestrator Test Orchestratorhuman owner Director, Quality Engineering routes into c2l

Estate rating

BBB

73 / 100

autonomy 86%cap 92%

Test generation, selection and triage are the highest-autonomy work in Engineering. The quality gate signature before release stays with a named human.

Agents

8

Lines

3

Units / wk

16,000

Timetable vs actual

6.9h → 6.6h

Cost / unit

$4.20

was $13

Run-rate saving

$7.4M

annualized at this volume

The 8 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.

Test Orchestrator

Owns the test & assurance lane end to end. Sequences the other agents, holds the timetable, and decides what surfaces to a human.

ObserveA3102 / 24h92%7% ovr$4.79/task

ceiling A4 · escalates to Escalates to Director, Quality Engineering when the unit falls outside the decision envelope or confidence drops below the floor.

AI GBSRuns the shared transaction spine.

Test Generation Agent

Runs the repeatable portion of test & assurance inside the shared transaction spine at spine cost and spine controls.

AI GBSA44,151 / 24h99%9% ovr$0.61/task

ceiling A4 · escalates to Escalates to Director, Quality Engineering when the unit falls outside the decision envelope or confidence drops below the floor.

Performance Test Agent

Runs the repeatable portion of test & assurance inside the shared transaction spine at spine cost and spine controls.

AI GBSA41,538 / 24h96%8% ovr$0.25/task

ceiling A4 · escalates to Escalates to Director, Quality Engineering when the unit falls outside the decision envelope or confidence drops below the floor.

TaskExecutes stations on a line.

Regression Selection Agent

Executes the test & assurance step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA42,826 / 24h94%3% ovr$0.47/task

ceiling A4 · escalates to Escalates to Director, Quality Engineering when the unit falls outside the decision envelope or confidence drops below the floor.

Flaky Test Agent

Executes the test & assurance step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA42,277 / 24h98%9% ovr$0.10/task

ceiling A4 · escalates to Escalates to Director, Quality Engineering when the unit falls outside the decision envelope or confidence drops below the floor.

Test Data Agent

Executes the test & assurance step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA43,964 / 24h94%9% ovr$0.34/task

ceiling A4 · escalates to Escalates to Director, Quality Engineering when the unit falls outside the decision envelope or confidence drops below the floor.

Coverage Gap Agent

Executes the test & assurance step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA41,211 / 24h96%7% ovr$0.35/task

ceiling A4 · escalates to Escalates to Director, Quality Engineering when the unit falls outside the decision envelope or confidence drops below the floor.

ChallengerPaid to disagree before a human has to.

Test & Assurance Challenger Agent

Adversarial reviewer for test & assurance. Argues the opposite case on every position and flags where the evidence does not carry the claim.

StrategyA252 / 24h95%5% ovr$10/task

ceiling A4 · 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 Coverage75
Evidence Quality79
Override Discipline87
Data Integrity60
Recovery Readiness72
Cost Transparency96
Judgment Readiness47

People

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

headcount on this work84.554.9 FTE

Roles held flat; released hours absorbed volume growth rather than headcount reduction.

inference cost per unit$0.78

Lines running in this sub-function

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

8 clear8 evidenced1 held
EN5ARelease Test CycleTransactional · 6 stations

Select regression set

1 live

Generate new tests

23 q

Execute

1 live

Triage failures

1 live

Escalate blocker

2 q

Sign quality gate

1 live

EN5BFlaky Test RemediationTransactional · 5 stations

Detect flake

1 live

Diagnose

12 q

Quarantine

2 live

Fix

46 q

Restore

2 live

EN5CPerformance BaselineTransactional · 4 stations

Run load

1 live

Compare baseline

9 q

Flag regression

4 live

File issue

3 live

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.

EN5-POS-100

Test & Assurance Operating Position

Rejectedv3

Which parts of test & assurance should deliberately stay human, and how do we defend that against a cost argument?

Hold the ceiling at A4 for test & assurance. Test generation, selection and triage are the highest-autonomy work in Engineering. The quality gate signature before release stays with a named human. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.

confidence

88%

The challenge — Test & Assurance Challenger Agent

Test & Assurance 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 13 percent for two consecutive months
  • ·Cost per test run stops falling while volume keeps rising
  • ·A control failure in this lane reaches a customer or a regulator

Evidence gaps

  • ·No external comparator on a like-for-like unit definition
  • ·Exception cases under $42k are sampled, not fully measured
Alternative consideredCostRiskVerdict
Hold the current position$585k run rateKnown and priced. Cedes ground if comparators move faster.live
Raise the ceiling one level now$1165k to build controlsOverride rate is 5 percent; raising the ceiling before that settles imports the error into production.rejected on evidence
Move the exception tail to the spine$265k transitionLoses local context. Rework risk on the cases that are hardest to recover.under review
authored by Test Orchestratorratifier Director, Quality Engineering3 stated assumptions

Policy register

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

RefRuleScopeOwner agentApproved byVersionStatus
EN5-POL-200

Test & Assurance Decision Envelope

Agents in this lane may act without a human when the test run 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 · Global

envelopeTest OrchestratorDirector, Quality Engineeringv1.7Active
EN5-POL-201

Test & Assurance 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

evidenceTest OrchestratorDirector, Quality Engineeringv3.4Active