LensReading which lens this session carries.

Closing the loop

How far the written rules actually reach

620 of 1,361 agents in the estate are named by no policy rule at all. That was an observation. It is now 6 tracked targets with a named owner, a number and a date, measured live from the records every time this page is opened. 3 of them are enforced by a gate rather than reported.

Tracked targets
6
4 named owners between them
Measured live
6
every target runs a query
At or above target
0
every measured target is short
Enforced by a gate
3
the rest report a number and refuse nothing
Agents already past the gate
204
acting alone with no rule naming them

Every reach target, furthest behind first

The definition is what is being counted, the measurement is taken now, and the gap is the distance to the number somebody signed up to.

Failed positions with a consequence recorded
failure-consequence
The share of positions whose outcome check says the position did not hold and which carry a learning record naming something that changed as a result.
reported only · No gate. Nothing stops a position failing quietly and nothing changing.
A learning record that says nothing changed does not count toward this figure.
30.0%
6 of 20
target 100.0% by 2027-03-31
Avery Chen
Enterprise Chief of Staff
70.0 points short
Committed forecasts compared on time
forecast-comparison
The share of forecasts attached to a committed scenario that were compared against an outcome on or before the date written on them.
gate active · A scenario cannot be committed without a forecast, and a forecast cannot be written without a comparison date. Both are enforced in the data model - the columns do not accept a blank.
Counted against forecasts, not scenarios. One committed scenario carries all of them.
60.0%
3 of 5
target 100.0% by 2026-12-30
Ana Duarte
Group Financial Controller
40.0 points short
Checks signed outside the authoring team
check-independence
The share of verification checks on the record whose signer sits in a different team from the one that authored the thing being checked.
gate active · A self-check may not be counted toward an independence figure. The platform enforces that: it stores the two teams separately and compares them. It cannot make the second person exist.
A check counts as independent when the checking team differs from the authoring team.
24.7%
62 of 251
target 60.0% by 2027-06-30
Grace Abbott
Head of Internal Audit
35.3 points short
Agents named by at least one rule
agent-rule-cover
The share of agents in the estate whose name appears in the binding list of at least one policy item, whatever the state of that policy.
reported only · No gate. This one is a target and nothing enforces it.
Counted by matching every name in every policy binding list against the agent estate.
54.4%
741 of 1,361
target 85.0% by 2027-06-30
Avery Chen
Enterprise Chief of Staff
30.6 points short
Rule cover for agents that act alone
autonomous-rule-cover
The share of agents at autonomy A3 or A4 - the levels at which an agent acts without a person in the loop - whose name appears in the binding list of at least one policy item.
gate active · An agent may not be set to A3 or A4 unless at least one policy names it. The platform can refuse to count an unnamed agent as compliant and can list every one that is already past the line. It does not provision agents, so it cannot stop one being created.
A3 and A4 are the levels at which an agent acts without a person in the loop.
74.7%
601 of 805
target 100.0% by 2027-03-31
Grace Abbott
Head of Internal Audit
25.3 points short
Break conditions that can actually be tested
break-evaluability
The share of authored break conditions that map to an expression the platform can run against data it holds.
reported only · No gate. A position can still be ratified with break conditions that nothing can test.
Evaluable means the sentence maps to an expression the platform can run against data it holds.
33.3%
209 of 627
target 50.0% by 2026-12-30
Morgan Idris
Function Leader
16.7 points short
The distance is arithmetic on the records behind each target. It is not a plan, and nothing on this page assigns any of it to anybody. The one target that now has a plan behind it

Agents acting alone with no rule naming them

The gate says a new agent cannot reach autonomous status without a rule that binds it. These are the ones already past it, which the platform lists rather than relaxing the gate.

Account Planning Orchestrator
Account Planning
A3 · active
Advocacy Orchestrator
Customer Advocacy
A3 · active
Advocate Health Agent
Customer Advocacy
A3 · active
Anomaly Detection Agent
Attribution & Analytics
A3 · active
Anomaly Detection Agent
Management Reporting
A4 · active
Application Health Agent
Application Management
A3 · review
Application Management Orchestrator
Application Management
A3 · active
Asset Orchestrator
Asset & License
A3 · active
Audit Request Agent
Internal Controls & Audit Interlock
A3 · active
Automation Factory Orchestrator
Automation Factory
A3 · active
Background & Right-to-Work Agent
Talent Acquisition
A3 · active
Bias Detection Agent
Experiment Design
A3 · active
Billing Interlock Orchestrator
Billing Interlock
A3 · active
Budget Orchestrator
Budgeting & Forecasting
A3 · active
Budget Pacing Agent
Demand Generation
A4 · active
Build Orchestrator
Build
A3 · active
Campaign Orchestrator
Campaign Orchestration
A3 · active
Campaign Quality Agent
Campaign Orchestration
A4 · active
Candidate Comms Agent
Talent Acquisition
A3 · active
Case Management Orchestrator
Case Management
A3 · active
Catalog & Intake Orchestrator
Service Catalog & Intake
A3 · active
Change Orchestrator
Change & Release
A3 · review
Channel Orchestrator
Partner & Channel
A3 · active
Churn Signal Agent
Renewals & Churn Ops
A4 · active

Actions

What is waiting on a person

Reach is a number with an owner and a date against it, not a hope. These are the ones behind, furthest behind first.

Operations

What this desk is allowed to start

A surface that only reports is not operable. This is the work this page can set in motion, and the bound it runs into.

Trigger and bound

This surface turns policy reach into a tracked metric: a definition, an owner, a target, a date, and a query that measures it now. Where a gate is active it refuses to let an agent past a status threshold without a rule that binds it. It does not write the missing rules, and it does not choose who owns the gap.

Live observability

What the record shows right now

Every reach target the platform holds, split 1 ways by where it stands against its own number. The single bar is 6.

Current distribution

6 targets

Short of target6100%

Is policy and strategy coming to fruition

Whether the written intent is holding here

No. Reach is now measured and owned, and every measured target is below its number.

Not holding on the record

The written intent is that policy reach stops being an observation and becomes a thing somebody is accountable for. 6 targets carry a definition, an owner and a date; 6 are measured by a live query; 3 are enforced by a gate rather than reported. 4 people own them. What has moved is the accountability. What has not moved is the reach itself - 620 agents are still named by no rule.

Where these numbers come from

Every figure on this page is measured live from the records at the moment the page is drawn. Nothing is stored. A target has an owner and a date; a gate is a rule the platform can test and report against. Where a gate is already being crossed, the agents past it are listed rather than the gate being quietly relaxed.