LensReading which lens this session carries.

Function cockpit

Engineering

Agents build, test and document. Engineers own architecture and release.

Autonomy0% / 80%

Phase: Phase 3 — Functional orchestration

Run the intent-to-release graph as a governed system: decomposition, code generation, testing, documentation, vulnerability detection and deployment preparation are agent work; product intent, architecture and production release remain human.

Trust score

0

Touchless

0%

Human review

0%

Override rate

0%

Capacity released

25-40% engineering capacity in selected software-factory settings (case-specific)

Accountable human

Chief Technology Officer

CoS agent: Engineering Chief of Staff

Eleven decomposed sub-functions, each with its own estate rating

Sub-functions

Engineering, 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

70%92%

The binding constraint here is Architecture & Design — capped at 70%.

Agents author the design options and the decision record. Architecture decisions constrain years of future work, so a review board ratifies them.

Portfolio & Intake

Engineering Intake Orchestrator · human owner VP Engineering Operations

BBB

estate rating

Owns request work end to end inside Engineering. Runs 3 lines across 16 stations, drains the repeatable share into the C2L tower, and holds 4 human gates.

autonomy now 67%ceiling 74%

Triage, estimation and dependency mapping are agentic. What the engineering organization spends its quarter on is a leadership commitment.

Units / wk

4,462

Agents

8

Lines

3

Cost / unit

$15

was $42

1 Orchestrator2 AI GBS3 Task1 Strategy1 Challenger c2l
5 clear3 evidenced2 held
10 live units
Architecture & Design

Architecture Orchestrator · human owner Chief Architect

BBB

estate rating

Owns design decision work end to end inside Engineering. Runs 3 lines across 17 stations, drains the repeatable share into no shared tower, and holds 3 human gates.

autonomy now 66%ceiling 70%

Agents author the design options and the decision record. Architecture decisions constrain years of future work, so a review board ratifies them.

Units / wk

6,946

Agents

8

Lines

3

Cost / unit

$8.80

was $21

1 Orchestrator4 Task1 Policy1 Strategy1 Challenger
6 clear4 evidenced0 held
10 live units
Build

Build Orchestrator · human owner Director, Platform Engineering

A

estate rating

Owns change work end to end inside Engineering. Runs 3 lines across 16 stations, drains the repeatable share into no shared tower, and holds 4 human gates.

autonomy now 76%ceiling 86%

Code generation and dependency maintenance are near-touchless. A human still approves every merge into a protected branch — that gate is the whole control.

Units / wk

15,527

Agents

8

Lines

3

Cost / unit

$15

was $40

1 Orchestrator6 Task1 Challenger
12 clear4 evidenced3 held
19 live units
Code Review & Quality

Code Quality Orchestrator · human owner Director, Engineering Quality

BBB

estate rating

Owns pull request work end to end inside Engineering. Runs 3 lines across 14 stations, drains the repeatable share into no shared tower, and holds 5 human gates.

autonomy now 82%ceiling 88%

Analysis and comment generation are fully agentic and never get tired. Trivial changes auto-approve; anything touching security or data paths requires a human reviewer.

Units / wk

9,574

Agents

8

Lines

3

Cost / unit

$14

was $47

1 Orchestrator4 Task1 Observe1 Policy1 Challenger
12 clear4 evidenced4 held
20 live units
Test & Assurance

Test Orchestrator · human owner Director, Quality Engineering

BBB

estate rating

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.

autonomy now 86%ceiling 92%

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

Units / wk

16,000

Agents

8

Lines

3

Cost / unit

$4.20

was $13

1 Orchestrator2 AI GBS4 Task1 Challenger c2l
8 clear8 evidenced1 held
17 live units
Release & Deployment

Release Orchestrator · human owner Director, Release Engineering

A

estate rating

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

autonomy now 81%ceiling 85%

Canary analysis and automatic revert are faster than any human and run unsupervised. Promotion to full production passes a change board.

Units / wk

14,063

Agents

8

Lines

3

Cost / unit

$6.74

was $21

1 Orchestrator2 AI GBS3 Task1 Policy1 Challenger it
16 clear3 evidenced1 held
20 live units
Site Reliability

Reliability Orchestrator · human owner Director, Site Reliability

A

estate rating

Owns incident work end to end inside Engineering. Runs 3 lines across 17 stations, drains the repeatable share into the AI-OPS tower, and holds 3 human gates.

autonomy now 82%ceiling 87%

Correlation and runbook execution run in seconds without a human. A severity-one incident always gets a named human commander — the agent supports, it does not command.

Units / wk

16,109

Agents

9

Lines

3

Cost / unit

$6.59

was $24

1 Orchestrator2 AI GBS3 Task2 Policy1 Challenger ai-ops
10 clear5 evidenced1 held
16 live units
Technical Debt

Technical Debt Orchestrator · human owner Director, Engineering Effectiveness

A

estate rating

Owns debt item work end to end inside Engineering. Runs 3 lines across 16 stations, drains the repeatable share into no shared tower, and holds 3 human gates.

autonomy now 71%ceiling 84%

Detection, migration and deprecation are mechanical and agentic. Accepting the remediation into a team backlog is an owner decision.

Units / wk

12,503

Agents

8

Lines

3

Cost / unit

$17

was $43

1 Orchestrator1 Observe4 Task1 Strategy1 Challenger
13 clear2 evidenced2 held
17 live units
Developer Experience

Developer Experience Orchestrator · human owner Director, Developer Platform

A

estate rating

Owns developer request work end to end inside Engineering. Runs 3 lines across 16 stations, drains the repeatable share into the AI-OPS tower, and holds 2 human gates.

autonomy now 85%ceiling 90%

Environment provisioning and toolchain setup are fully self-service and touchless. Only platform roadmap choices need a human.

Units / wk

13,955

Agents

8

Lines

3

Cost / unit

$13

was $36

1 Orchestrator2 AI GBS2 Task1 Observe1 Strategy1 Challenger ai-ops
13 clear3 evidenced2 held
18 live units
Security Engineering

Security Engineering Orchestrator · human owner Director, Security Engineering

A

estate rating

Owns finding work end to end inside Engineering. Runs 3 lines across 19 stations, drains the repeatable share into the IT tower, and holds 4 human gates.

autonomy now 72%ceiling 82%

Scanning, credential revocation and low-risk patching are automatic because speed is the control. Critical vulnerabilities always get a named human owner with a clock.

Units / wk

7,570

Agents

8

Lines

3

Cost / unit

$13

was $47

1 Orchestrator2 AI GBS2 Task2 Policy1 Challenger it
11 clear5 evidenced3 held
19 live units
Engineering Analytics

Engineering Analytics Orchestrator · human owner Director, Engineering Analytics

BBB

estate rating

Owns analysis work end to end inside Engineering. Runs 3 lines across 15 stations, drains the repeatable share into the DATA-ANALYTICS tower, and holds 3 human gates.

autonomy now 63%ceiling 88%

Metric assembly and trend detection are fully agentic. Team-level metrics are deliberately never used for individual performance without a human interpreting context.

Units / wk

139

Agents

8

Lines

3

Cost / unit

$441

was $2K

1 Orchestrator2 AI GBS2 Task1 Observe1 Strategy1 Challenger data-analytics
3 clear1 evidenced4 held
8 live units

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.

AAAAAABBBBBBCCCD