Three structural readings, each with its own evidence.

Complexity, Agility and RecoverabilityRecoverabilityThe R in C/A/R: how well a team or service restores acceptable service after a failure. Recovery times and tested alternatives can supply evidence. (C/A/RC/A/RThree separate scores for a team or service: Complexity, Agility and Recoverability. Each uses current evidence and the customer’s stated scoring rules.) are three separate measures of a team or service as it is now. Complexity is how much structure the unit has to manage: its systems, and the changes that have to be coordinated across them. Agility is how readily it can change safely: change lead timesLead timeThe elapsed time between starting a request or handoff and completing it. and change failures. Recoverability is how well it restores acceptable service after a failure: restoration times and tested recovery.

The three are kept apart on purpose. A unit can be simple to run and slow to change, or quick to change and slow to recover. A single blended number would hide that difference, so NovusLens does not produce one.

C/A/R is a measure of current structure, refreshed as evidence arrives. It is not a prediction of how a team will perform, and it is not a judgement of individuals. The three measures differ from the accountability questions of CapabilityCapabilityWhat a team can provide or an agent can do. NovusLens records the claim and the evidence that supports it., AuthorityAuthorityWhat a team or agent is allowed to decide or do. For an agent, this comes from a named person or system that grants permission. and ResponsibilityResponsibilityThe work or result that a team or agent answers for. NovusLens records the person answerable for it.. These describe what an AI agent may do and who answers for it. The two share initials only. The accountability questions are always written out in full and are covered on agentic accountability.

What a score is made of

A raw figure such as a count of systems or a restoration time is not a score. A score exists only where three things are in place.

The first is a calibrationCalibrationThe customer's own rule that turns raw figures, such as a gauge, a workflow history or an incident log, into a C/A/R score. A score moves when the calibration moves, and the change is recorded.: the agreed rule that converts observations into an indicator. The organisation writes it. A Recoverability rule might start from 100 and subtract 5 points per hour of restoration time. The example below works through it. A calibration names the unit, the measure and the scopeScopeThe work, data, activity or access that a rule, statement or calculation covers. For access, it names the records an account can see. For team measures, “Own operations” covers the selected unit. “Including descendants” adds the units beneath it.. It then gives a starting value and the inputs used to adjust it. A document must support the calibration, at the shownShownEvidence that a document supports a value. A recovery plan that states a target shows the target, but it does not show that the target was achieved. rungRungOne of three levels of support for a recorded value: told, shown or measured. The level appears beside the value. on the evidence ladder. A figure with no calibration behind it is not calculated.

The second is scope. “Own operations” uses only the selected unit. “Including descendants” also uses every unit beneath it that is owned by it. External suppliers stay outside the ownership sum. A parent’s figure is recalculated from the records within its scope. It is never an average of its children’s scores, because an average would blur the records that the figure rests on.

The third is current evidenceCurrent evidenceA view that uses only measurements recent enough to count. The “What we were told” view uses recorded statements instead.. Observations carry a time window. Once a window closes, the observation no longer counts toward the current figure.

An example

Take an illustrative calibration for Recoverability, with a base of 100 and one term: total restoration hours over the period, at minus 5 points per hour. For 6 restoration hours inside the window, the calculation is 100 minus 30, giving 70. NovusLens keeps the result between 0 and 100. The organisation chose the base and the rate, and the rule is visible. If the calibration is changed, the score moves on the record. The tool does not choose a weight of its own.

If the same unit also has a calibration for Agility but no current change records, its Agility is unknown. It is not zero.

How to read it

Read each measure on its own, and read the scope beside it. A score of 70 has no meaning across units unless the calibrations, scopes and observation periods match. Open the calculation to see the calibration, the records behind each term, the dates, and whether a higher value is favourable.

A figure can have one of three states. Available means one current calibrated claim supports it. Unknown means no current calibrated evidence exists for that scope. Reasons include an expired calibration, a changed scope or missing required records. Conflict means more than one current claim exists and they have not been reconciled.

Unknown is not zero. A missing figure means the input is missing, conflicting or out of date. It does not mean the unit has no complexity, or that its recovery is perfect.

Some sources can measure only one measure. For example, an incident feed can measure recoverability but not complexity. By default a score is written only when all three measures are available. A calibration can set partial_axes, which writes each measure that can be measured and leaves the others unknown. Any measure that cannot be measured is reported as unknown, with the reason.

A key-person alert

The mapMapThe view of teams, services and their dependencies for the selected part of the organisation. Details appear beside the map. also raises a key-person riskKey-person riskA finding about a team on the measured critical path with one to four people represented in the records. It never names those people. findingFindingA difference in the records that needs a decision. It includes its sources and can describe a slow handoff, disputed fact or missing owner.. It fires when a cellCellA box on the map that stands for a team, service, supplier or AI agent. It is a unit you can inspect, not necessarily a department or a person. sits on the measured critical pathCritical pathThe dependency chain with the greatest total coordination weight in the map. It identifies a structural constraint. It does not state elapsed time. and has one to four people represented in the current personnel mapping. A cell with no mapped people does not raise a key-person finding. Missing ownership stated in an interview can raise unowned-capability separately. The privacy floorPrivacy floorThe rule that a unit with fewer than five mapped people is shown as a small team, never as a number. The key-person finding says fewer than five and never says who. is five. Below that number the exact count is withheld, because in a very small team the count identifies the person. The finding names where the exposure sits and does not name anyone.

The critical path is measured. The person count is not: it is what the personnel mapping documents about who is assigned to the cell today. A count of one is a structural exposure worth raising. It is not proof that work depends on that person, and a count of ten is not proof of resilience. The finding is an alert for the accountable ownerAccountable ownerThe person recorded as answerable for a team, service or agent. NovusLens does not assume that this person is also the one who runs it day to day. to resolve, not advice on hiring.

Limits

C/A/R measures structure, not performance. It cannot tell you how well a team is run.

The calibration is an organisation’s own claim. Writing it down makes the assumptions attributable, so they can be challenged. It does not make them correct. The reading is therefore only as good as the calibration behind it.

When internal dependenciesDependencyA link where one unit needs something from another. A dependency is not a reporting line: it shows where work relies on something else. form a cycleDependency cycleA loop in the dependencies: one unit needs another, which eventually needs the first unit again. For example, Sales waits on Finance and Finance waits on Sales., NovusLens reports a dependency-cycle finding and does not infer a complexity score. Overlapping or conflicting records are excluded and listed, not silently added together.

What to do next

Open the unknown or conflicting figure first, and read its reason. Ask the measurement owner whether the missing input exists and whether the calibration still describes the unit. Then read the score next to the critical path for the same scope, because the two use the same dependency structure. To see what the product will and will not do with these figures, read the Layer-A/B boundary.

Where it is implemented

• apps/graph/src/conway_graph/car_assessment.py selects the included units and required records, and decides when to withhold a score. Its scores() function applies calibrations and returns available, unknown or conflict for each measure.
• apps/reasoning/src/conway_reasoning/car.py is the read path. It checks which records the reader may see and refuses a selected unit outside that scope. It returns an unknown figure if any unit below it is hidden.
• apps/connectors/src/conway_connectors/car_axes.py writes per-measure observations from calibrated sources and leaves uncalibrated measures unwritten.
• apps/graph/src/conway_graph/keyperson.py raises the key-person risk finding and holds the privacy floor of five.