Risk
Authority and consequence
Systems that can move money, change records, communicate externally, or affect people require stronger controls and evidence.
Commercial model
Pricing is confirmed after public-safe discovery because agent count is not a reliable measure of system complexity. Authority, data, tools, integration, evidence, and failure exposure drive scope.
Planning ranges
The theme intentionally does not encode binding prices. Commercial terms are confirmed in a written scope.
| Engagement | Typical duration | Commercial shape | Acceptance focus |
|---|---|---|---|
| Machine Intelligence Readiness Sprint | About 3 weeks | Fixed scope after discovery | Decision package and release conditions |
| Agentic Swarm Architecture Blueprint | About 4–6 weeks | Fixed scope after discovery | Approved architecture and sequence |
| Swarm Evaluation & Release Gates | About 6–8 weeks | Fixed scope or phased implementation | Reproducible evaluation and release gates |
| Agent Control Plane Implementation | About 8–12 weeks | Milestone-based implementation | Working controls for one bounded workflow |
| Machine Intelligence Operations Office | 3–6 month initial term | Defined monthly cadence | Recurring decisions and operating records |
| Joint Discovery Workshop | 2 working days plus preparation | Fixed workshop | Next-scope recommendation or no-go |
Scope drivers
These factors matter more than the number of prompts or agents.
Risk
Systems that can move money, change records, communicate externally, or affect people require stronger controls and evidence.
Access
Sensitive data, cross-tenant context, external tools, and legacy integrations expand architecture and test coverage.
Evidence
Reconstructing state, messages, model versions, tool results, approvals, and outcomes may require new instrumentation.
Operations
Availability, latency, provider constraints, records, incident response, and procurement shape the solution.
Commercial discipline
A proposal should state the decision, system boundary, artifacts, acceptance, dependencies, exclusions, and change process.
Start with a bounded decision
A public-safe conversation is used to identify the smallest engagement that can produce a credible decision.