LensReading which lens this session carries.

Standard operating procedure · GB7A · AI GBS & Global Business Services

Idea to Production

Transactional procedure owned by Automation Factory in AI GBS & Global Business Services. 9 numbered steps, 2 of them a person deciding and 7 of them agent work …

Procedure code

GB7A

Transactional

Rule versions in force

v3.4 · v2.2

2 governing rules stacked over it

In force since

2025-11-07

Earliest effective date across those rules

Last reviewed

2026-07-12

Latest review date across those rules

Rules approved by

1

Head of Automation

Rule health

1/2

1 under review · 0 superseded

There is no document register behind this header. The version, the dates and the approver names are read off the governing rules this procedure runs inside, which is the closest thing the estate holds to a document control record. Nobody wrote these procedures. There is no document library behind this page and no Word file anywhere in the estate. Every procedure below is generated from rows this platform already holds: the line and its ordered stations, the owning team and the person accountable for it, the policy items bound to that team, and the governance gates seeded against the function. The steps, the owners, the gates, the approvers, the dwell times and the volumes are read straight off those rows. The sentences that join them together are generated. Read a procedure as the estate describing itself, not as a document somebody signed.

Purpose and scope

What this procedure is for and where it stops

To run idea to production to one repeatable standard inside Automation Factory, part of AI GBS & Global Business Services.

Entry condition

Runs transactional. Starts when an upstream system posts an event.

Exit condition

The unit clears monitor.

In scope

  • Every unit of idea to production raised inside Automation Factory, from the trigger below through to the final step.
  • 7 of the 9 steps, which run without a person present.
  • 2 decisions that stop the line until a named person releases it.

Out of scope

  • Anything outside the envelope set by the owning team. It leaves the line at a gate and becomes the named approver's decision.
  • Work that belongs to another function. This line records no handoff, so nothing leaves it mid-flow.
  • The 1 step carried on the shared AI GBS spine. They are done for this line but not by this function.

Who owns it

Runs the line

Automation Factory Orchestrator

Accountable person

Head of Automation

The end of the escalation path. Above them sits GBS Chief of Staff for the function.

That team’s remit

Owns automation work end to end inside AI GBS & Global Business Services.

That remit describes the whole team, not this one procedure, so its counts are larger than the ones on this page.

Autonomy

68% now · ceiling 80%

Build and test are agentic.

The procedure, step by step

9 numbered steps · 2 of them stop for a person

A step is one station on the line.

No.StepWho does itWhat happensSystem of recordTime on the stepIn flight
01

Mine opportunity

Agentworkflow

Opportunity Mining Agent

Opportunity Mining Agent performs this step unattended at the workflow layer, inside the envelope its owning team has set, and writes an evidence record naming the actor, the authority it acted under and the digests of what it read and wrote.Enterprise Data Warehouse3.0 hrsstandard 4.5 hrs30
02

Score feasibility

Agentworkflow

Feasibility Scoring Agent

Feasibility Scoring Agent performs this step unattended at the workflow layer, inside the envelope its owning team has set, and writes an evidence record naming the actor, the authority it acted under and the digests of what it read and wrote.Group ERP — Financials1.4 hrsstandard 3.3 hrs40
03

Business case

Agentworkflow

Build Agent

Build Agent performs this step unattended at the workflow layer, inside the envelope its owning team has set, and writes an evidence record naming the actor, the authority it acted under and the digests of what it read and wrote.Core HR and Payroll2.1 hrsstandard 4.3 hrs32
04

Approve

HumanpolicyDecision

Head of Automation

Releases: Head of Automation

Work stops. Head of Automation reviews what the line has produced and either releases the unit or sends it back. Nothing downstream of this step runs until that decision is made.Holds until: Held for a named human when the unit exceeds the automation decision envelope, when confidence falls below the floor, or when a policy clause is engaged.Group ERP — Financials3.2 hrsstandard 7.7 hrs4
05

Build

Agentworkflow

Deployment Agent

Deployment Agent performs this step unattended at the workflow layer, inside the envelope its owning team has set, and writes an evidence record naming the actor, the authority it acted under and the digests of what it read and wrote.Source-to-Pay Suite2.2 hrsstandard 3.6 hrs21
06

Test

Agentworkflow

Regression Monitor Agent

Regression Monitor Agent performs this step unattended at the workflow layer, inside the envelope its owning team has set, and writes an evidence record naming the actor, the authority it acted under and the digests of what it read and wrote.Invoice Capture and AP Automation50 minstandard 1.1 hrs22
07

Control review

HumanpolicyDecision

Head of Automation

Releases: Head of Automation

Work stops. Head of Automation reviews what the line has produced and either releases the unit or sends it back. Nothing downstream of this step runs until that decision is made.Holds until: Held for a named human when the unit exceeds the automation decision envelope, when confidence falls below the floor, or when a policy clause is engaged.Group ERP — Financials53 minstandard 1.7 hrs9
08

Deploy

GBSgbs

Test Harness Agent

Test Harness Agent performs this step unattended on the shared AI GBS transaction spine, outside the function, inside the envelope its owning team has set, and writes an evidence record naming the actor, the authority it acted under and the digests of what it read and wrote.Group ERP — Financials1.2 hrsstandard 1.4 hrs36
09

Monitor

Agentworkflow

Build Policy Agent

Build Policy Agent performs this step unattended at the workflow layer, inside the envelope its owning team has set, and writes an evidence record naming the actor, the authority it acted under and the digests of what it read and wrote.Planning and Consolidation30 minstandard 54 min37

The swimlane

The same steps, laid out by who is holding the work

Three lanes: a person deciding, an agent working, and the shared services spine.

Person2 stations and 27% of the dwellAgent6 stations and 65% of the dwellShared spine1 station and 8% of the dwell5 crossings between lanes
Human
decides
Agent
does the work
GBS
shared spine
System
attributed by rule
Enterprise Data Warehouse
Group ERP — Financials
Core HR and Payroll
Group ERP — Financials
Source-to-Pay Suite
Invoice Capture and AP Automation
Group ERP — Financials
Group ERP — Financials
Planning and Consolidation
Click any station to see who does it, what it costs in time, the rule that gates it and the system attributed to it.

Run the procedure

One unit walked along that swimlane on the happy viable agentified path

Press run and a single unit enters at station 01. Each station takes exactly the dwell this platform records for it and is held by whoever the estate says holds it.

speed
worked clock
00:00
allowance clock
00:00
stations cleared
0/9
the happy path, worked time0 min of 15 hrs
the same run against its allowance0 min of 29 hrs
Human
decides · 2
Agent
does the work · 6
GBS
shared spine · 1
run log

Nothing has happened yet.

where the worked time goes15 hrs a unit
Person4.1 hrs · 27%Agent9.9 hrs · 65%Shared spine1.2 hrs · 8%5 crossings between lanes
what makes this path viable
  • The lanes are the recorded lanes. 2 of the 9 stations stop for a person and every one of them still does on this run.
  • Every one of the 2 decisions is still taken by the people who hold them. The happy path is that each one is released first time, not that it is skipped.
  • The owning team is held to a 80% autonomy ceiling and this run sits at 73% unattended, inside it.
  • 6 systems are attributed across the stations, and 0 of them are connected. Nothing in this run has been watched executing against a real system of record.
what would have to change

Nothing moves. 73% of the worked time already runs unattended and the owning team is held to a 80% ceiling, so the next decision cannot be handed over until somebody raises that ceiling.

stays with a person
  • 04 Approve — Handing this over would put 107% of the worked time in the unattended lane, past the 80% ceiling this team is held to.
  • 07 Control review — The last decision on the line. Something on every procedure ends with a person, so this one is never offered.
what this run is, and what it is not

This is a run, not a forecast.

Happy means nothing goes wrong: no station raises an exception, no gate sends the unit back, no rework loop fires and nothing sits in a queue between stations.

The second clock is the allowance.

A decision may be handed to the agent lane only if two things hold …

The run reads the station rows only.

What this run does not model: arrival patterns, contention between procedures for the same person, shift patterns, weekends, learning curves, and the cost of anything.

Run it again, going wrong

The same unit, the same stations, broken the way the estate already records breaking

This is the same unit on the same stations, breaking.

No break below was chosen.

One play head, two tracks, the same scale. The clean run stops when it runs out of stations; the other one keeps going.

As written

9 stations, nothing waiting, nothing sent back

0 min of 15 hrs

now Mine opportunity · Opportunity Mining Agent · 3.0 hrs

As the ledger has it

the same stations, plus 3 categories the estate already counts

0 min of 31 hrs

now Mine opportunity · Opportunity Mining Agent · 3.0 hrs

personagentshared spinework the break addsnobody working it

What breaks, and where it is counted

No break below was chosen.

What the two runs say together

3 counted categories land on this procedure.

What this run cannot claim

These breaks are shown together on one unit so the shape of a bad day fits on one screen.

Two clocks run. Worked time is the sum of the time the work is actually being done …

Roles and responsibilities

9 named parties

The responsibility grid is assigned by a rule, not lifted from a chart.

PartyRACIStepsDutyEscalates to

Automation Factory Orchestrator

orchestrator

Runs the line0Sequences this line, holds its timetable and decides what has to surface to a person. Raises to Head of Automation when a unit falls outside the envelope or confidence drops below the floor.

Opportunity Mining Agent

agent

A3% autonomy · ceiling A3%

86% clean · 13% overridden

Responsible1Watches automation factory continuously. Escalates to Head of Automation when the unit falls outside the decision envelope or confidence drops below the floor.

Feasibility Scoring Agent

agent

A3% autonomy · ceiling A3%

96% clean · 4% overridden

Responsible1Executes the automation factory step it owns, posts its evidence, and hands the unit to the next stage.Escalates to Head of Automation when the unit falls outside the decision envelope or confidence drops below the floor.

Build Agent

agent

A3% autonomy · ceiling A3%

90% clean · 2% overridden

Responsible1Executes the automation factory step it owns, posts its evidence, and hands the unit to the next stage.Escalates to Head of Automation when the unit falls outside the decision envelope or confidence drops below the floor.

Head of Automation

person

Accountable and responsible2Holds 2 decisions on this line. Accountable to Head of Automation for the decisions held here.

Deployment Agent

agent

A3% autonomy · ceiling A3%

92% clean · 4% overridden

Responsible1Executes the automation factory step it owns, posts its evidence, and hands the unit to the next stage.Escalates to Head of Automation when the unit falls outside the decision envelope or confidence drops below the floor.

Regression Monitor Agent

agent

A3% autonomy · ceiling A3%

86% clean · 16% overridden

Responsible1Watches automation factory continuously. Escalates to Head of Automation when the unit falls outside the decision envelope or confidence drops below the floor.

Test Harness Agent

agent

A3% autonomy · ceiling A3%

98% clean · 5% overridden

Responsible1Runs the repeatable portion of automation factory inside the shared transaction spine at spine cost and spine controls.Escalates to Head of Automation when the unit falls outside the decision envelope or confidence drops below the floor.

Build Policy Agent

agent

A2% autonomy · ceiling A3%

90% clean · 15% overridden

Responsible1Holds the rules that bind every other agent in automation factory. Escalates on any decision that would breach a live policy version, with the clause cited.

Decision points and escalation

2 places this procedure stops

04 · Approve

3.2 hrs

Held for a named human when the unit exceeds the automation decision envelope, when confidence falls below the floor, or when a policy clause is engaged.

Head of Automation

07 · Control review

53 min

Held for a named human when the unit exceeds the automation decision envelope, when confidence falls below the floor, or when a policy clause is engaged.

Head of Automation

Handoffs out of this procedure

Nothing leaves this procedure for another function. It completes inside Automation Factory.

The rules this procedure runs inside

2 governing rules

Automation Factory Decision Envelope

under review

GB7-POL-200 · envelope · vv3.4

Agents in this lane may act without a human when the automation sits inside the stated value, risk and confidence envelope.

Approved by Head of Automation · in force 2025-11-07 · reviewed 2026-06-21

Binds Feasibility Scoring Agent, Build Agent, Test Harness Agent, Deployment Agent, Benefit Realization Agent

LATAM · US

Automation Factory Evidence Standard

active

GB7-POL-201 · evidence · vv2.2

Every autonomous decision writes inputs, the rule version applied, the model and prompt version, the output and a reversal path.

Approved by Head of Automation · in force 2025-12-16 · reviewed 2026-07-12

Binds Feasibility Scoring Agent, Build Agent, Test Harness Agent, Deployment Agent, Benefit Realization Agent

EU · UK · US

Systems this procedure touches

6 named · 0 connected

System attribution is a rule, not a reading.

SystemVendorDomainCriticalityStepsConnection
Group ERP — FinancialsSAP S/4HANAledgercritical4modeled
Enterprise Data WarehouseSnowflakeanalyticscritical1modeled
Core HR and PayrollWorkday HCMhcmcritical1modeled
Source-to-Pay SuiteSAP Aribaprocurementcritical1modeled
Invoice Capture and AP AutomationBaswareprocurementhigh1modeled
Planning and ConsolidationAnaplanledgerhigh1modeled

0 of the 6 systems on this procedure is connected to the platform, and none of the twenty-six in the estate register is. 5 of them are marked as taking a write back, which is a property of the system, not a live path from here.

How it performs

End to end

15.3 hrs

Summed from the time each step sits, standard 28.5 hrs — running at 54% of the standard.

Touchless

73%

Agent share

78%

A week

6,593

In flight

231

Effort a week

246 human minutes · 671 agent minutes

Volume multiplied by the time on each step.

Assurance over this procedure

4 governance gates sit above it

The twenty-four controls and eighteen exception classes in this platform are held at estate level …

GateWhat trips itWho releases itClockWaivable

Autonomy level increase

approval

Any move of an agent or workflow to a higher autonomy bandBusiness owner and risk5 business daysNo

New data class processing

approval

Any service processing a data class not already approvedRisk and privacy officer5 business daysNo

Outsourcing or service removal

approval

Any material operating-model, sourcing or service-portfolio changeGBS executive committee10 business daysNo

Agent recertification

oversight

Quarterly evaluation of every agent against its published certification criteriaAgent governance boardQuarterlyNo

What this document is and is not

These procedures describe the modeled estate.

Actions

What is waiting on a person

What one procedure document gives you, and what to check before you treat it as the standard.

Operations

What this desk is allowed to start

A surface that only reports is not operable.

Trigger and bound

A procedure document prints purpose, scope, entry and exit, the numbered steps with the actor on each, the roles and their RACI position, the swimlane, the decision points and escalation, the handoffs, the governing rules with version and approver, the systems by rule and the timings.

Live observability

What the record shows right now

The whole library, by lane, so a single document can be read against the estate it came from.

Current distribution

2,748 steps

Agent, inside the function1,38951%
Agent, on the shared spine85131%
Person deciding at a gate50818%

Is policy and strategy coming to fruition

Whether the written intent is holding here

Half. The document is complete and every actor on it is measured

Not holding on the record

A procedure here carries a version, an approver, an in-force date and a last-review date.