Straight answers about bolde.

Three kinds. Frontier cloud models from selected providers. Private models: open-source or open-weight models run locally in the selected environment. Sovereign models: owned by the customer, either brought in or created and tailored by ASG under an agreed scope.

Hosting is chosen separately from the model. ASG-hosted bolde. runs in U.S. data centers with SOC 2 controls. bolde. can also run in a customer cloud environment, Azure, AWS or on-premises, following the selected provider, region and configuration.

Per deployment, in writing. Your level (Open, Hybrid, Private or Sovereign) is written into your contract in the Security Level Schedule.

Yes. A deployment can use cloud models, private models or both, chosen per task. A frontier cloud route is selected by separate agreement, and only content approved to leave the deployment boundary reaches it.

Yes. Start where the work is today and change level when it needs more control or less. A move is a dated change to your Security Level Schedule, agreed before it happens.

There is no single model. Each deployment records which model is selected, who owns it, where it runs and which routes it may use.

Your data is yours. You get it back when you leave. You also own the business content your team produces with bolde.

Industry models are never trained on client data. If your deployment includes a model tuned on your work, your tuned model is never pooled with anyone else’s and never trains anyone else’s.

Where your hosting choice and security level say, and nowhere else. Both are written into your contract.

Every answer shows its source. Analyst links each figure to the source table, query and row range behind it, so you can check the answer against the original record.

Auditor keeps a tamper-evident record of queries, answers, citations and approvals. Each entry references the product, data source and output hash. Audit packages can be exported for your team to review.

Access is set for each deployment and each source. Lake shows who can use each item. Space owners grant view or edit access to one artifact or a whole Space.

The systems your team already uses: email, documents, CRM, ERP, data warehouses and other common business systems. Lake uses existing connectors where they fit.

bolde. can connect to it through a custom connection, built as part of your bolde. implementation. New connections use authorized APIs, exports, databases or other approved interfaces, with access and scope agreed for the deployment.

No. Keep what you run. bolde. connects to it.

No. Lake does not connect to a source without your approval, and Analyst does not read data sources you did not configure.

ASG’s Deployment Engineers. They connect bolde. to your systems, ship it to production and own those connections after go-live. Agentic Secure Group (ASG) is the AI platform company that makes bolde.

Phase 0 of your bolde. deployment. We configure bolde. for your workflows and security level, and confirm the Build scope. Your timeline and go-live date are set in Phase 0.

One bolde. order form covers Phase 0, the Build and your bolde. subscription. The subscription starts at go-live.

We price phases, not hours. bolde. implementation is a fixed fee per phase, set on your order form. Subscriptions are $10K, $15K or $25K per month, locked at the rung you join.

Your bolde. Navigator handles calls and support. The Deployment Engineer who built each connection keeps owning it after go-live.

We only take on work that we deliver on bolde.

No. We don’t place people. We deploy bolde.

Request early access and tell us the first workflow you want to put to work. One workflow is enough to start.

Pilots are paid. Each starts with one workflow, at the security level set for that deployment. You expand once that workflow proves itself.

Your timeline and go-live date are set in Phase 0, before the Build starts.

Sign up for bolde. Updates to get product notes and release updates. The roadmap page shows what is planned next.