One repeated decision, one approved outcome, one accountable owner, and one written evidence bar.
One named workflow. One scoped quote. No seat or usage meter.
Nodes prices each decision program as one workflow from trigger through approved execution and evidence. The quote reflects the workflow boundary, systems, operating envelope, deployment controls, implementation, and support.
No public list price · No per-seat meter · No token or model-call meter · Quote changes only when scope materially changes
One decision. 72 hours. A written $1M evidence bar.
Start with one repeated high-stakes decision and at least two years of linked decisions and outcomes. Nodes confirms whether the history has enough volume, usable join keys, outcome coverage, data quality, and policy overlap to support the test.
The clock begins only after the agreed deidentified dataset is delivered under a mutual NDA and accepted. The customer's finance team defines the $1M threshold and its calculation in writing.
Read-only historical analysis after Nodes accepts the agreed deidentified dataset. No live decision is touched.
Receive the evidence, limitations, and whether the historical record supports the finance-approved threshold.
Decision Replay identifies counterfactual or historical evidence. It does not establish future realized value.
If the history does not support the written bar, there is no production fee.
The customer keeps the analysis, including the limitations. The remedy applies to the production deployment fee and is not a published product price or a refund claim.
License the approved workflow from trigger through outcome.
The production quote covers the work the decision program must perform. Approved team members can use the licensed workflow without a per-user charge.
A defined event starts the program.
The customer fixes which event, system, population, and operating condition can initiate the workflow.
Nodes reads approved context in place.
Connectors gather only the evidence permitted by the Decision Blueprint and deployment boundary.
The outcome engine returns a decision and drafted action.
The review includes rationale, evidence, limitations, expected business effect, and systems affected.
A named human controls consequential execution.
The accountable owner questions, edits, approves, delays, or declines the proposed action.
Agents and workflows execute the approved steps.
The orchestrator coordinates permitted work across the systems inside the licensed scope.
The Decision Trace records the complete path.
Evidence, versions, human input, approvals, and execution remain queryable under customer policy.
The downstream outcome improves the next decision.
The defined result joins the customer context graph and the ongoing calibration record.
Workflow-based pricing. No internal-usage meter.
The proposal ties commercial terms to delivered work instead of consumption inside the approved program.
What is not metered
- Seats, users, candidates, or applicants
- Tokens or model calls
- Recommendations, approvals, or actions
- Software credits
What sets the quote
- Workflow and approved outcome
- Systems, data, and integrations
- Operating envelope and controls
- Deployment, implementation, and support
Prove the decision first. License the workflow second.
Payback begins with the customer's economics. Finance approves the inputs, measurement method, guardrails, and evidence bar before production terms are scoped.
The customer retains the customer-owned assets defined in the agreement, including data, learned weights, Blueprints, Traces, outcome history, and calibration artifacts. The license, agent operation, connector maintenance, monitoring, updates, support, and continuous recalibration end with the service.
Start with the decision.
Bring one repeated high-stakes decision, at least two years of linked outcomes, and the finance owner who can define the evidence bar.