Why a private LLM matters for your business.
Public AI tools work by sending your inputs to a shared model. For most businesses that is fine. For healthcare practices, RIAs, broker-dealers, law firms, and government contractors, it is a compliance event waiting to happen. AiTLLM puts an Intelligent IT-managed control plane between your users and the model, so your data does not move without your knowledge, your consent, and your audit trail attached.
Compliance, built in from day one
An audit pack ready for HIPAA, FINRA, SOX, GDPR, and CMMC reviewers. Prompt logs, model versions, human-review records, and data lineage are captured automatically. When your auditor asks, you hand them a document, not a screenshot.
Your data stays in a controlled boundary
No egress to public model providers. Prompts, embeddings, fine-tune data, and outputs stay inside an Intelligent IT-governed perimeter. The logs you can show your auditor are the same logs that run your system.
One IT partner for everything
AiTLLM sits inside your existing Microsoft 365 or Google Workspace tenant. SSO, SharePoint and Drive RAG connectors, DLP policies, and line-of-business app integrations are handled by the same team managing your endpoints and your security stack.
What Intelligent IT operates for you.
We run the full stack. Your team uses AI; they do not learn ML ops, manage model endpoints, or write DLP policies from scratch.
- Model routing and selection. We pick the right model for your workload, route between providers, and swap when something materially better ships. You do not track the open-weight model landscape; we do.
- RAG over your existing data. SharePoint, Google Drive, Confluence, Notion, ticketing systems, CRM. Your team queries your data in plain English; we wire the retrieval pipeline and keep it fresh.
- Identity and tenant isolation. Email-domain-based tenant assignment. SSO via Entra ID or Google Workspace. No cross-tenant leakage. Per-department personas and prompt libraries.
- Audit trail and DLP. Every prompt is logged with user identity, model version, timestamp, and redaction status before it leaves your perimeter. DLP filters run at the gateway, not as an afterthought.
- Evidence packs on demand. Pre-built templates for HIPAA, SOC 2, FINRA 17a-4, GDPR, and EU AI Act Article 50. Pulled automatically into the AiT Trust Portal and available to your compliance team at any time.
- Ongoing operations, not just deployment. The same Intelligent IT engineer who manages your endpoints and MDR is the named owner of your AI gateway. One call, one relationship, one contract.
Three tiers. One decision.
Start where you are. Move up when you are ready. All tiers share the same Intelligent IT control plane; the infrastructure model and the level of isolation change between them.
AiTLLM Connect
Connect
For small teams who want to use ChatGPT-class AI without sending client data through a consumer chatbot.
Plus token passthrough at cost + 15%. No setup fee. Month-to-month.
Bring your own Claude, OpenAI, or Gemini API key.
- Up to 25 named users
- 1 RAG connector (SharePoint, Drive, Confluence, or Notion)
- Multi-tenant gateway with email-domain routing
- SSO via Google + Microsoft
- DLP filter rules and prompt logging
- Monthly evidence pack (HIPAA / GDPR / SOC 2)
- Named Intelligent IT technical contact
AiTLLM Private
Private
For regulated mid-market firms who want a real private LLM without standing up a data-science team.
Per tenant pod (flat) + $39 per active user / mo.
$4,500 one-time onboarding. Includes dedicated model infrastructure.
- Up to 200 active users
- Up to 5 RAG connectors
- Isolated tenant pod on Together AI, Azure PTU, Bedrock, or Vertex
- Llama 4, Qwen 3.6, DeepSeek V3 / R1, Kimi K2.6, GLM-5.1, Mistral, or Claude routing
- Custom prompt library and per-department personas
- HIPAA + SOC 2 + GDPR evidence pack on AiT Trust Portal
- Quarterly model review and migration
- Named account engineer + 8x5 support, 4-hour P1 response
AiTLLM Sovereign
Sovereign
For regulated firms with a real reason to own the silicon. Defense subcontractors, hospital systems, broker-dealers.
Retainer + actuals (hardware, cloud, license).
$25,000 one-time setup. Customer VPC or on-prem. Air-gap variant available.
- Unlimited users, unlimited RAG sources
- Architecture, deploy, and hardening inside your perimeter
- Identity integration (Entra ID, Okta, Google Workspace)
- AiT gateway, RAG, audit, and DLP on customer infra
- Quarterly DR drill and model-update window
- 24x7 incident response with named engineer
- Annual third-party audit support (SOC 2, HIPAA, FedRAMP-aligned)
Compliance and audit coverage.
AiTLLM ships with pre-built evidence templates for the frameworks your auditor actually cares about. These are not checkbox exercises; they are the logs your system generates by running.
HIPAA
Business Associate Agreement, PHI non-egress controls, audit log retention, and access-control evidence for covered entities and their business associates.
FINRA / SEC Rule 17a-4
Prompt and output retention on WORM-compliant storage, model-version attribution, and human-review records for broker-dealers and RIAs from day one.
SOX
AI-generated content flagged for human review in financial reporting workflows. Change-management logs for model swaps. Separation-of-duties evidence.
GDPR / EU AI Act Article 50
Data residency controls, right-to-erasure workflow, prompt-marking for AI-generated content, and transparency obligations for the August 2026 enforcement deadline.
CMMC Level 2
CUI boundary enforcement at the gateway, access-control alignment with NIST SP 800-171, and audit log continuity for defense contractors in the CMMC certification cycle.
ISO 27001 / SOC 2 Type II
AI gateway included in the information security management system scope. Model inventory, change records, and incident logs pulled into the AiT Trust Portal for continuous evidence.
Deliverables for every tier at Connect and above: audit log export, data lineage report, model usage attribution summary, vendor risk artifact for your third-party risk register.
US AI compliance, May 2026
The audit pack regulators and insurers actually want.
The Colorado AI Act was delayed and paused, but Texas TRAIGA is in force with civil penalties up to $200,000 per violation. California AB 2013 took effect January 1, 2026. NYC Local Law 144 enforcement is tightening after a damaging State Comptroller audit. SEC, FINRA, NYDFS, FTC, HHS OCR, and the OCC/FRB/FDIC have all named AI as a 2026 examination priority. Cyber-insurance carriers added AI questionnaires to renewal applications in early 2026 and started writing AI-specific exclusions.
Every regime asks for the same artifacts: an AI inventory, a NIST AI RMF self-assessment, impact assessments per high-risk use case, vendor BAAs and DPAs, consumer notices, retention policies, and AI-aware incident response. AiTLLM Private and Sovereign tiers ship that audit pack pre-built so the SMB’s day-one posture matches what an examiner expects on day ninety.
How it works.
Three steps from prompt to response. The Intelligent IT control plane sits in the middle and handles routing, RAG retrieval, identity scoping, DLP enforcement, and audit logging before anything reaches the model or returns to your user.
Your users
M365 / Workspace / custom UI
Intelligent IT Control Plane
Your data never leaves this boundary unlogged
Partner-hosted model
Together AI, Azure PTU, Bedrock, Vertex, or on-prem
Response + audit record
Every output timestamped, attributed, and logged
AiTLLM as the substrate for the AiT product family.
AiTLLM is not a standalone chatbot. It is the AI layer that powers every other product Intelligent IT ships. If you are already using any AiT product, you are already running on the AiTLLM gateway.
- AiTCRM. Client-summary generation, meeting-prep drafts, and deal-intelligence pulled from the CRM record and your connected knowledge sources.
- AiTBMS. Natural-language query over WebCTRL and BACnet telemetry. Ask a building in plain English what happened to HVAC zone 4 last Tuesday.
- AiT SOC Sentinel. Analyst-facing triage summaries with redacted PII. The LLM helps the human analyst focus; it does not replace the analyst.
- AiT Trust Portal. Compliance narrative drafts generated from your live evidence pack. The model writes the first draft; your compliance team reviews and signs.
FAQ
Start with a 30-minute discovery call.
We walk through your data environment, your compliance requirements, and which tier fits your current stage. No slide deck. No sales engineering queue.
Page as of 2026-05-07. Pricing reflects current AiTLLM tier reference points; final scope confirmed in discovery. Token costs pass through at vendor list + 15% on Connect. Manuel Ruiz, Founder / CEO / CISO. © Intelligent Group (DBA Intelligent IT) · intelligentit.io