aml

AMM · Model

Ageni Measure Model

Count a number once. Trust it everywhere it shows up.

Request access

Why a number means one thing

A measure is only as trustworthy as the specification it is typed against. These are the three bodies AMM answers to, and what each one makes true about a figure you read.

SMM — OMG

Structured Metrics Metamodel, Object Management Group

SMM types every measure. A measure's type is not a label we picked — it is the category the specification puts it in, which is why two people counting the same thing arrive at the same KIND of number before they ever compare values.

FIBO — EDM Council

Financial Industry Business Ontology, EDM Council

FIBO grounds the financial measures. A revenue figure resolves to the ontology's concept of revenue rather than to a column heading, so the number means the same thing to your controller, your lender and your board.

ISO / IEC

International Organization for Standardization

ISO/IEC sits behind the classification. The vocabulary a measure is classified against is an international standard rather than house style, so a measure keeps its meaning outside our walls — including in an audit.

One true number for everything you count.

Every figure in your business means one thing, computed one way — so two reports never disagree.

Same fact, two different numbers.

Revenue shows up one way on the P&L and another way on the forecast. Margin means one thing on a dashboard and something else in a spreadsheet. They're the same fact, spelled differently. AMM folds every derivation onto one measure, so the two reports can't disagree.

before

$67,645 revenue. Margin 55%. GM 0.55. Three spellings, one fact.

in ageni

Every derivation flows into one measure genome.

after

One identity. Every report reads the same number.

The law most models get wrong

industry and industry-per-sector are different measures

Not the same measure with a filter on it. A scope restriction produces a distinct measure identity. Revenue and expense ventilate across dozens of segments — channel, catalogue, product-per-catalogue, sector, industry-per-sector, language, region — and recursive X-per-Y scoping is law across all of them.

Any comparative benchmark that cannot tell those two apart is reporting on the wrong peer set — and it will look completely normal while it does.

Recursive segment scoping

A scope restriction mints a new measure identity — it does not filter an existing one. Compose a scope chain and watch the measure change into a different measure.

Identity amm-measure-collective-tally-of-gross-revenue-stream-across-monthly-window
Genome facet segment_scope []
Peer set

Three cases that must never be conflated
Segment-scoped absolute
The measurand narrowed to one segment coordinate — a filter, not a division. Spelled …-within-{dimension}-segment. Spelling it with by would mis-parse a filter as a ratio, so that form is rejected here.
Segment share
The scoped absolute divided by the unscoped whole — a genuine proportion, and the lever segments are ranked by. Spelled ratio-proportion-of-{scoped}-by-{whole}.
Segment average
The scoped tally divided by the count of members in the segment — a genuine average. Spelled collective-average-of-{scoped}-by-{member-count}.

Segment-scoped absolute

a filter, not a division

The base measurand narrowed to the subset belonging to one segment coordinate — gross revenue attributed to one product catalogue. Its shape is a sum over a filtered stream.

Segment share

a genuine proportion in [0,1]

The scoped absolute divided by the unscoped whole — one catalogue's share of total gross profit. This is the lever segments are ranked by.

Segment average

a genuine average

The scoped tally divided by the count of members in the segment — revenue per product within a catalogue.

Conflating them is a naming failure before it is a reporting failure. A scope restriction spelled as a division makes the name mis-parse as a ratio, so AMM gives scope its own connective — within, closing on the reserved word segment — and never overloads the division operand. Scopes nest, innermost first:

amm-measure-collective-tally-of-gross-revenue-stream-across-monthly-window-within-catalogue-segment-within-sector-segment

The name spells the shape. Read it back and you know exactly which population was counted, over which window, at which depth of scope — without opening the query.

1,155
nodes on the live tree
743
measures, deduped by signature
404
verified causal edges

What Ageni Measure Model gets you

Every one of these is a real outcome the engine helps a buyer reach — not a feature list.

One true number, everywhere you look.

↑ reliability

Never have two reports disagree.

↑ reliability

Stop reconciling numbers by hand.

↓ time

Trust every figure without double-checking.

↑ risk

Measure if a change actually worked.

↑ speed

Compare any two numbers safely.

↑ reliability

Ground your numbers in a finance standard.

↑ risk

Cut time lost to number disputes.

↓ cost

See your numbers as one connected tree.

↑ speed

Roll every metric up to profit.

↑ speed

Keep the same meaning across every app.

↑ reliability

Report to anyone with confidence.

↑ reliability

Grounded in

OMG — Object Management GroupEDM Council

AMM is the Universal Measure Model. Every number in the business is forced onto a base alphabet of 15 SMM-typed measures, the OMG Structured Metrics Metamodel, and a KPI tree grounded in FIBO, the Financial Industry Business Ontology. A metric and a KPI are not new measures; they are roles layered on the same computed genome, so two derivations of one quantity always fold back to a single identity.

  1. 1 Forces every business number onto 15 SMM-typed base measures
  2. 2 A KPI tree grounded in FIBO, the Financial Industry Business Ontology
  3. 3 A computed genome gives one identity per measure, never duplicated
  4. 4 metric and kpi are roles on that genome, not new measures

AMM is in private beta. Request access for your organization.

Request access