Workflows
Deterministic automations for documents, approvals, extraction, routing, and reporting.
Automations
Workflows, Team Members, and Sims priced by useful workload.
The automation story is explicit: Business includes fair use for everyday work, AI coworkers, and customer simulation loops. Heavier jobs draw from usage balance, and repeatable enterprise workflows can graduate into customer-owned local models.
Agents
Task-oriented workers that research, draft, validate, and coordinate across tools and data sources.
Team Members
Persistent virtual employees for recurring internal work, role-specific execution, and reusable business knowledge.
Sims
Customer or target personas used to test plans, content, offers, and agent behavior before those flows reach real users.
Pricing logic
A middle ground between mystery and noise.
Automation pricing is explained in buyer language: simple workloads are included or cheap, heavy jobs are metered, and recurring enterprise Team Members or Sims can become strategic assets instead of perpetual variable cost.
| Workload type | Typical pattern | Commercial treatment |
|---|---|---|
| Light text automation | Short prompts, templated outputs, approval loops | Business fair use |
| Deep research | Multi-source retrieval, ranking, validation, citations | Fair use plus usage balance for heavier jobs |
| Sim review loop | Persona-based plan review, objection handling, content feedback | Fair use plus usage balance when workloads become heavy or continuous |
| Media automation | Image, music, or video jobs triggered by workflows | Usage balance |
| Persistent Team Member program | Recurring enterprise process with stable workload and owned knowledge | Custom Enterprise packaging and owned-model path |
Caching
Repeated context, recurring Team Member behavior, and repeated answers do not need to be recomputed at full price every time.
Batching
Jobs can be grouped and sent more efficiently, especially for document workflows, reporting pipelines, and simulation runs.
Endpoint optimization
Different endpoints, models, and pricing tiers can be selected to improve cost without giving up privacy or deployment requirements.