The RaaS Architecture
The RaaS architecture connects resolution definition, instrumentation, attribution, economics, contracting, and the AI Trust Layer into one commercially defensible system.
Resolution as a Service is not a pricing change applied to an otherwise unchanged SaaS product. It is an interdependent product, financial, contractual, and verification architecture.
A vendor cannot price a resolution until it has defined the unit of completed work. It cannot measure contribution margin until it can identify the full cost of delivering and verifying that resolution. It cannot defend an invoice until the completion evidence, attribution rules, exclusions, and dispute process are contractually usable.
The RaaS architecture has six connected components.
1. Resolution Definition
The vendor must define the discrete business problem being resolved, the triggering state, the required completion state, the scope boundary, and the conditions that make the event billable.
A mature definition also identifies exclusions, duplicate-counting rules, service-level windows, human intervention, reopen treatment, reversal conditions, and the authoritative system of record.
See Resolution Definition Maturity.
2. Instrumentation and Resolution Evidence
The product and surrounding systems must produce evidence sufficient to determine whether the defined resolution occurred.
Evidence may come from customer-controlled records, systems of record, auditable event logs, workflow state changes, third-party telemetry, and independent verification services. Vendor telemetry may be part of the evidence set, but should not automatically be treated as the sole authority.
3. Attribution and Causation
A completed business event is not automatically a vendor-caused resolution.
The architecture must distinguish work materially performed by the vendor’s platform from work completed by a human, another software system, an external process, or a combination of contributors. Attribution rules should be explicit enough to prevent the vendor from billing for participation when the contract requires causation.
4. Resolution Economics
Each resolution class must have a measurable revenue and cost profile.
Resolution Contribution Margin should include the costs directly attributable to delivery: inference, retrieval, orchestration, human oversight, quality assurance, verification, exception handling, failure, rework, and dispute support.
A resolution can produce customer value and still be commercially unsustainable if its cost-to-serve is not understood and controlled.
5. Contract Architecture
The contract must translate the operational definition into enforceable commercial rules.
It should state what counts as a resolution, what evidence controls, how attribution is determined, which conditions disqualify billing, how reopens and reversals are treated, how duplicate counting is prevented, and how disputes are adjudicated.
The commercial structure may combine a platform fee, included resolution volume, per-resolution overage pricing, exposure caps, and renegotiation thresholds. The precise structure can vary, but the billable unit and evidence rules cannot remain ambiguous.
6. The AI Trust Layer
The AI Trust Layer provides the independent verification, evidence, attribution, and governance infrastructure required to make resolution claims credible beyond the vendor’s own reporting environment.
It supports invoicing, audit, procurement, insurance, service-level enforcement, financial controls, and institutional acceptance. Without it, the vendor may deliver the outcome, classify the event, count it, and issue the bill without independent validation.
Why the Components Must Be Designed Together
These components fail when implemented independently:
- A precise resolution definition without instrumentation cannot be proven.
- Instrumentation without attribution cannot establish causation.
- Attribution without exception handling can overcount reopened or reversed work.
- Pricing without full delivery and verification cost can create negative-margin growth.
- Contract language without reviewable evidence cannot survive a dispute.
- Vendor-controlled evidence without independent validation creates a structural conflict of interest.
The architecture is complete only when the resolution can be defined, delivered, measured, attributed, verified, priced, contracted, audited, and governed as one coherent system.
RaaS changes what software vendors are paid for. The AI Trust Layer determines whether those claims can be verified, contracted, audited, and trusted.