The package combines purpose-built local AI computing with a ready-to-use enterprise AI experience. Lenovo ThinkStation PGX supplies the NVIDIA Grace Blackwell–based hardware and AI software foundation. iMBrace adds preconfigured AI assistants, Knowledge Drive folders, a browser-based user experience, access controls, installation guidance, training, and a defined support path. This helps organizations move from owning AI-capable hardware to using it for everyday knowledge work.

PGX is the local hardware AI computing platform; iMBrace is the enterprise software application and enablement layer. Every configuration includes five prebuilt SME business assistants, one iMBrace Reference Assistant, preconfigured Knowledge folders, document-grounded search, centralized access permissions, and secure browser extension integration. Instead of requiring customers to assemble a complex application stack themselves, this unified solution provides a fully unlocked, ready-to-use foundation for private business AI on Day 1.

A standalone PGX gives technical teams a powerful environment for AI development and inference. Building a usable internal service still requires software selection, interface development, knowledge organization, access design, testing, user training, and ongoing support. iMBrace packages these business-facing capabilities into a pre-configured, turnkey solution (starting with our single-node AI Essentials configuration) so organizations can deploy immediately on Day 1 and scale physical capacity as their needs grow.

No. iMBrace makes selected AI capabilities accessible to business users through a browser experience and preconfigured assistants. Users can ask questions about approved company, product, sales, finance, procurement, or reference materials without working directly with notebooks or AI development frameworks. An authorized administrator or reseller manages installation, access, and approved content.

The solution can process approved business knowledge within a customer-managed PGX deployment, helping reduce dependence on public cloud inference for supported workloads. The exact data flow depends on the models, integrations, remote access, telemetry, and services enabled for the deployment. Customers should review the final architecture and security configuration with their reseller or iMBrace before go-live. The package should not be described as air-gapped unless that configuration has been specifically designed and approved.

The built-in knowledge experience retrieves information from documents placed in approved Knowledge Drive folders; uploading a document does not by itself mean training a public foundation model. Any model training, fine-tuning, external connector, or third-party AI service would require separate configuration and review. Data treatment should be confirmed in the applicable agreement and deployment architecture.

No. Generative AI can produce incomplete or incorrect answers, especially when source documents are unclear, outdated, conflicting, poorly extracted, or outside the assistant’s assigned knowledge. The assistants should support—not replace—professional judgment. High-impact decisions should remain subject to the organization’s established review and approval process.

The iMBrace platform shortens preparation by providing a standardized, rapid installation process, preconfigured assistants, ready-to-use Knowledge folders, and comprehensive onboarding materials. Actual readiness depends on hardware delivery, network and domain preparation, administrator availability, approved business documents, and completion of customer acceptance testing. Your participating partner will confirm the implementation schedule.

iMBrace supports common business formats such as PDF, Microsoft Word, Excel, PowerPoint, CSV, and common image formats. The iMBrace software does not impose a separate per-file upload size limit; the total amount of content that can be stored is subject only to the physical SSD storage capacity of your selected configuration (e.g., 2TB on a single node). Processing results and speed also depend on file format, document quality, and complexity, which can be verified during acceptance testing.

Yes. Our single-node AI Essentials configuration provides a fully capable, flexible starting point to experience prebuilt AI assistants and validate your workflows. As your data footprint and user base grow, you can seamlessly expand by adding physical PGX hardware nodes (scaling to AI Team or AI Scale configurations) and upgrading your SLA support levels. Because the full software platform is unlocked standard from Day 1, you never have to worry about locked software features or paying for software license upgrades to access advanced tools.