FAQ
Everything you need to know about private AI deployment, the audit process, pricing, compliance, and working with Northline Systems.
Private AI infrastructure means running AI models — like large language models (LLMs) — on hardware you physically own and control, inside your own office or data center. Your data never leaves your building. This is the opposite of cloud AI services like ChatGPT or Claude, where your data is processed on someone else's servers.
We deploy Apple Mac Mini M4 Pro systems with 48GB of unified memory. These machines are capable of running open-source AI models locally with excellent performance. Hardware is provided at cost (~$1,700) and you own it outright.
We deploy open-source models via Ollama — typically DeepSeek, Llama, Mistral, and Phi variants. Model selection depends on your use case and performance requirements. Sensitive tasks run locally; non-sensitive tasks can route to cloud APIs (Claude, GPT-4) for maximum quality through our hybrid routing layer.
Hybrid routing is our approach to getting the best of both worlds. A rules engine classifies each task: privileged documents, client data, and sensitive information stays on your local hardware. General research, public information queries, and non-sensitive drafting routes to cloud AI for superior quality. Your team gets one portal — the routing happens automatically.
Your team accesses a custom web portal from any browser on your office network. It looks and feels like ChatGPT but runs on your infrastructure. Team members can upload documents, ask questions, draft content, and run automations — all without data leaving your building.
A 3-business-day engagement where we assess your current AI exposure, classify your data, design your system architecture, write an AI usage policy for your team, and build a working prototype of your first automation. Cost is $3,500, fully credited toward your deployment build.
Customized intake questionnaire, 90-minute discovery session, written AI Operations Report, security assessment, data classification framework, AI usage policy, working prototype with live demo, Phase 1 build recommendation with exact pricing, and 30-day follow-up support.
Yes. 7-day money-back guarantee. If the deliverables don't meet your expectations, you get a full refund.
You keep everything — the report, the security assessment, the data classification framework, the AI usage policy, and the prototype. These deliverables have standalone value regardless of whether you deploy.
Yes. The standard audit is fully remote. We also offer an Audit + On-Site Assessment ($5,000) that includes a physical visit to evaluate your infrastructure, network, and hardware placement. The on-site option is recommended for businesses in the Coeur d'Alene and Spokane corridor.
The foundation platform starts at $18,000. This includes the Mac Mini deployment, open-source model configuration, security hardening, web portal, hybrid routing layer, and staff training. Additional modules are $5,000 each. AI Receptionist is $7,500. Typical first engagement: $26,000–$33,000 including modules. Hardware at cost (~$1,700). Full audit fee credited.
Managed services are $2,997/month. This includes monthly model updates, prompt tuning, system monitoring, security patching, and a performance report showing documents processed, calls handled, time saved, and recommendations. Approximately 10 hours of dedicated engineering per month.
No contracts. Managed services are month-to-month. You own all hardware, all code, and all customizations we build. If you cancel, everything stays with you.
Most clients see positive ROI within 2-4 months. A 12-attorney law firm recovering 25+ hours per week of admin time at $325/hr blended rate generates over $340K/year in recovered billable capacity — the deployment pays for itself many times over.
Yes. Because all protected health information (PHI) is processed on hardware you own inside your facility, no PHI is transmitted to external servers. This eliminates the primary compliance risk of cloud AI. We also provide a HIPAA-specific AI usage policy as part of the audit.
Yes. Financial advisors subject to SEC Regulation S-P and FINRA rules can use private AI infrastructure because client financial data never leaves the firm's controlled environment. The hybrid routing layer ensures regulated data stays local while non-sensitive tasks can use cloud APIs.
Yes. For government contractors handling Controlled Unclassified Information (CUI), on-premise AI deployment eliminates the risk of CUI being processed by unauthorized cloud providers. We configure systems to meet CMMC Level 2 requirements for AI tool usage.
Private AI deployment preserves attorney-client privilege because privileged communications and documents are never sent to third-party servers. Multiple state bar associations have issued guidance requiring firms to evaluate AI tools for privilege implications — on-premise deployment satisfies this requirement.
Docker sandboxing, firewall rules restricting external access, audit logging of all AI interactions, credential isolation, encrypted storage, and network segmentation. The system runs on your local network and is not accessible from the internet.
Coeur d'Alene, Idaho. We serve the entire Inland Northwest — Coeur d'Alene, Spokane, Post Falls, Hayden, Sandpoint, Moscow — plus clients across Idaho and the U.S. via remote engagements.
Kyle Cunningham, the founder. No account managers, no offshore teams, no subcontractors. The person who designs the architecture is the person who deploys the hardware and answers your phone calls.
Audit: ~3 business days. Build & deploy: 1-2 weeks. 14-day hypercare post-launch. Most clients go from first call to production system in under 30 days.
Yes. The audit and deployment work for organizations anywhere in the U.S. Remote model configuration, portal deployment, and managed services work the same regardless of location. On-site visits are available for businesses in Idaho and Eastern Washington, with travel available for larger engagements.
Book a free 15-minute call. We'll answer your specific questions and determine if the AI Operations Audit makes sense for your business.