LensReading which lens this session carries.
Research & Development/Regulatory Science

Sub-function

Regulatory Science

Owns submission work end to end inside Research & Development. Runs 3 lines across 19 stations, drains the repeatable share into the C2L tower, and holds 6 human gates.

orchestrator Regulatory Science Orchestratorhuman owner VP Regulatory Science routes into c2l

Estate rating

BBB

72 / 100

autonomy 63%cap 66%

Requirement mapping and evidence assembly are agentic at scale. Everything filed with a regulator is signed by a named officer with personal accountability.

Agents

8

Lines

3

Units / wk

22,214

Timetable vs actual

8.9h → 9.0h

Cost / unit

$15

was $40

Run-rate saving

$28.7M

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.

Regulatory Science Orchestrator

Owns the regulatory science lane end to end. Sequences the other agents, holds the timetable, and decides what surfaces to a human.

ObserveA2169 / 24h90%9% ovr$4.70/task

ceiling A2 · escalates to Escalates to VP Regulatory Science when the unit falls outside the decision envelope or confidence drops below the floor.

AI GBSRuns the shared transaction spine.

Requirement Mapping Agent

Runs the repeatable portion of regulatory science inside the shared transaction spine at spine cost and spine controls.

AI GBSA21,215 / 24h91%2% ovr$0.46/task

ceiling A2 · escalates to Escalates to VP Regulatory Science when the unit falls outside the decision envelope or confidence drops below the floor.

TaskExecutes stations on a line.

Evidence Assembly Agent

Executes the regulatory science step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA23,337 / 24h93%7% ovr$0.07/task

ceiling A2 · escalates to Escalates to VP Regulatory Science when the unit falls outside the decision envelope or confidence drops below the floor.

Submission Drafting Agent

Executes the regulatory science step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA21,508 / 24h90%4% ovr$0.23/task

ceiling A2 · escalates to Escalates to VP Regulatory Science when the unit falls outside the decision envelope or confidence drops below the floor.

Testing Coordination Agent

Executes the regulatory science step it owns, posts its evidence, and hands the unit to the next stage.

WorkflowA2345 / 24h99%4% ovr$0.31/task

ceiling A2 · escalates to Escalates to VP Regulatory Science when the unit falls outside the decision envelope or confidence drops below the floor.

ObserveWatches signals and watches the agents.

Standards Watch Agent

Watches regulatory science continuously. Detects drift, quantifies it, and routes what matters without waiting for a reporting cycle.

ObserveA295 / 24h90%15% ovr$12/task

ceiling A2 · escalates to Escalates to VP Regulatory Science when the unit falls outside the decision envelope or confidence drops below the floor.

PolicyAuthors and versions the binding rules.

Regulatory Policy Agent

Holds the rules that bind every other agent in regulatory science. Versions them, tests them against live decisions, and blocks work that breaches them.

PolicyA235 / 24h87%15% ovr$5.40/task

ceiling A2 · 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.

Regulatory Science Challenger Agent

Adversarial reviewer for regulatory science. Argues the opposite case on every position and flags where the evidence does not carry the claim.

StrategyA247 / 24h96%6% ovr$13/task

ceiling A2 · 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 Coverage59
Evidence Quality70
Override Discipline80
Data Integrity86
Recovery Readiness78
Cost Transparency84
Judgment Readiness56

People

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

headcount on this work65.843.1 FTE

Redeployed into customer-facing escalation ownership.

inference cost per unit$3.46

Lines running in this sub-function

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

14 clear2 evidenced5 held
RD8ARegulatory SubmissionTransactional · 8 stations

Map requirements

1 live

Assemble evidence

1 live

Draft dossier

2 live

Scientist review

2 q

Counsel review

1 live

Officer sign

5 q

Submit

27 q

Track

24 q

RD8BStandards Change ResponseTransactional · 6 stations

Detect change

3 live

Assess impact

2 live

Map affected products

32 q

Draft plan

1 live

Owner accept

1 live

Execute

2 live

RD8CCompliance TestingTransactional · 5 stations

Schedule

45 q

Execute

3 live

Record

1 live

Non-conformance escalate

3 live

Certify

9 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.

RD8-POS-100

Regulatory Science Operating Position

Ratifiedv3

Which parts of regulatory science should deliberately stay human, and how do we defend that against a cost argument?

Hold the ceiling at A2 for regulatory science. Requirement mapping and evidence assembly are agentic at scale. Everything filed with a regulator is signed by a named officer with personal accountability. Raise it only after two consecutive quarters where the override rate stays under five percent and every override has a written cause.

confidence

68%

The challenge — Regulatory Science Challenger Agent

Regulatory Science Challenger Agent argues the recommendation leans on 5 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 9 percent for two consecutive months
  • ·Cost per submission 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 $35k are sampled, not fully measured
Alternative consideredCostRiskVerdict
Hold the current position$898k run rateKnown and priced. Cedes ground if comparators move faster.not selected
Raise the ceiling one level now$282k to build controlsOverride rate is 9 percent; raising the ceiling before that settles imports the error into production.rejected on evidence
Move the exception tail to the spine$419k transitionLoses local context. Rework risk on the cases that are hardest to recover.selected
authored by Regulatory Science Orchestratorratifier VP Regulatory Science3 stated assumptions
outcome held

Ratified 6 months ago. Measured since: cost per submission fell 16 percent against a forecast of 11 percent.

Policy register

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

RefRuleScopeOwner agentApproved byVersionStatus
RD8-POL-200

Regulatory Science Decision Envelope

Agents in this lane may act without a human when the submission 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 4 agents · APAC

envelopeRegulatory Policy AgentVP Regulatory Sciencev3.3Active
RD8-POL-201

Regulatory Science 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 4 agents · LATAM · US

evidenceRegulatory Policy AgentVP Regulatory Sciencev1.4Active