Skip to content

07 / Adoption and cost intelligence

One graph, two initial decision surfaces.

Adoption intelligence shows where AI is taking hold across teams and workflows. Cost intelligence shows where spend concentrates and which ownership paths are supported by evidence. Both remain bounded by attribution coverage and deployment evidence.

Question

How do adoption and cost intelligence support different buyers without creating separate versions of ownership?

Mechanism

How the mechanism works

One attributed usage record can be aggregated for an AI program decision or inspected for an owner and budget decision. The buyer, aggregation level, threshold, and action change. The underlying ownership evidence does not.

Illustrative synthetic example

Decision record

Shared attributed record

Mechanism . Describes how the product is designed to work.
Workload
support-routing-workload
Invocation
invocation.synthetic.0241
Service
support-router
Project or code owner
Customer Operations Platform
Identity
svc-support-router
Organization
Customer Experience Systems
Budget responsibility
CX Automation

Operational confidence

Operational confidence

0.91

Output state

Strongly inferred

At or above the 0.80 review threshold

Explanation
Service, deployment, identity, organization, and billing evidence support one leading path.
Time basis
Event-time, synthetic evaluation window
Source health
All enabled synthetic sources healthy
Action eligibility
Decision support

Evidence sources

  • Provider billing record
  • Deployment metadata
  • Service identity
  • Organization hierarchy

Alternative candidates

  • Support Experience team, retained as an alternative
  • Platform Infrastructure, runtime owner only

Truth domains

Observed . Directly present in a source record. Inferred . Derived from eligible evidence with uncertainty preserved. Event-time . The state that applied when the governed event occurred.
Output state
Strongly inferred
Time basis
Event-time, synthetic evaluation window
Source health
All enabled synthetic sources healthy
Action class
Decision support
Correction history
No correction event in this synthetic record
Illustrative synthetic example
01 Initial projection . An initial decision surface derived from the attribution graph. Validation pending . Customer evidence is still required to validate actionability.

Adoption intelligence

Question: Where is AI taking hold in real work?

Team and function
Customer Experience Systems
Workflow
Support routing
Governed usage versus licensed access
Compared at program and cohort level
Tooling trajectory
Directional view over the governed evaluation window
Adoption depth
Workflow embedded, occasional, or unavailable
Coverage
Capturable governed usage denominator retained
Source provenance
Attributed record and enabled source set
Cohort granularity
Minimum cohort rules apply
Unknown or unavailable views
Visible rather than rendered as zero

Adoption intelligence is cohort and program intelligence derived from governed attributed usage. It is not employee-performance analytics.

Illustrative synthetic example
02 Initial projection . An initial decision surface derived from the attribution graph.

Cost intelligence

Question: Where does spend concentrate, and who is accountable for it?

Workload
support-routing-workload
Service
support-router
Owner
Customer Operations Platform
Budget
CX Automation
Estimated versus reconciled cost
Truth domain retained per amount
Operational confidence
0.91 in the synthetic record
Unresolved spend
Kept separate from supported ownership
Review eligibility
Decision support; finance policy remains separate

Finance eligibility remains separate from output state and operational confidence.

Annotated visual

Inspect the path and its unresolved states

Two decision surfaces read one attributed usage record
Governed attributed usage
evidence, owner path, confidence, coverage, time basis
AdoptionProgram and cohort viewenablement and workflow planning
CostOwner and budget viewspend concentration and review
Both branches stop at the same coverage, source-health, and unresolved-state boundary.

In summary: one governed attributed usage record branches into an adoption intelligence view for program and cohort decisions and a cost intelligence view for owner and budget decisions. Both branches retain the same attribution coverage, source health, time basis, and unresolved states.

Concrete example

Apply the mechanism to one synthetic workload

Illustrative synthetic example

Support-routing workflow, read two ways

AI program leadership

Inspect whether governed workflow usage exists across the eligible cohort, where licensed access has not become captured usage, and which source gaps limit the view.

Decision: define an enablement hypothesis for validation.

FinOps

Inspect the supported owner and budget path, estimated versus reconciled cost, unresolved spend, and finance-review eligibility.

Decision: route supported and unresolved records into separate review paths.

Evidence and implementation boundary

What the current claim supports

The architecture supports both projections. Customer evidence is required to show that each produces useful, actionable decisions. Adoption actionability remains design-partner validation pending.

Validation begins with one shared attributed record and separate decision criteria. Adoption validation must show that program teams can take useful action without individual scoring. Cost validation must show that owner and budget evidence can support a bounded finance or optimization review. Design-partner deployments define this evidence boundary before either decision surface is enabled.

Known limitations

Unknown and unresolved states remain valid outcomes

  • 01Adoption intelligence is cohort and program intelligence, not employee-performance analytics.
  • 02Licensed access does not prove governed usage, and missing capture does not prove non-adoption.
  • 03A supported owner path does not automatically make spend eligible for chargeback or policy action.

Decision enabled

Choose the buyer, operating question, evidence threshold, and validation measure for the same attributed usage record.

The validation plan should name the buyer, decision, minimum coverage, source health, time basis, output states allowed, and evidence that would count as useful action.