Frequently asked questions

Our smallest project engagement is approximately $180K, which corresponds to a 10–12 week scoped pilot on a single workflow. Below that threshold we tend to recommend other firms — the overhead of standing up our team isn't fair to a smaller engagement.

Yes — for clearly scoped pilots and integrations, we issue fixed-price quotes. For ongoing operations or open-ended programs, we use either monthly retainers or outcome-based pricing. We'll recommend the model that aligns incentives best for the work in front of us.

Yes. We routinely work through enterprise procurement, master service agreements, and statement-of-work structures, including with public sector clients. Our standard terms have been negotiated by procurement teams at top-25 banks and Fortune 100 industrials.

Selectively, with established companies, alongside cash compensation. We don't do pure equity deals — we've found alignment is cleaner when we are paid for the work and our incentives don't depend on a separate exit narrative.

We use the model best suited to the task. In practice we are model-agnostic and operate across major frontier providers and open-weight models, with the choice driven by latency, accuracy, cost, and data residency requirements. We do not have a financial relationship with any model provider.

No. We deploy into your cloud, your VPC, your tools. Our orchestration layer is portable. We'll recommend a platform stack if asked, but we have no incentive to push you toward any particular vendor.

All deployments default to the customer's cloud and the customer's data residency region. Data does not leave your environment unless your contract explicitly authorizes it. PHI, PII, and regulated financial data are handled under controls reviewed annually by independent auditors.

Each engagement defines a specific accuracy contract — for example, "≥99.4% match against analyst review on a stratified sample of 500 cases per week." The metric is agreed before build, instrumented during build, and tracked weekly thereafter.

Every action is logged with a reasoning trace, every outbound system call is idempotent, and every workflow has a documented rollback. Material errors trigger an incident review within 48 hours, with root cause and remediation shared with your team.

You own the workflow design artifacts and any custom code written specifically for your operation. We retain ownership of our orchestration layer, evaluation tooling, and reusable adapters, which are licensed to you for the duration of the engagement.

We support graceful exit. Our standard contract includes a transition clause: we will train your team on the system over a 90-day period, transfer all documentation and runbooks, and remain on standby for an additional 60 days at a reduced rate.

We include up to 40 hours of audit support per year in every retainer. We have walked external auditors, internal audit teams, and regulators through our deployments. We provide pre-built evidence packs mapped to common control frameworks.