AICP Sovereign · Customer-controlled AI platform
Govern your models,
knowledge and agents.
AICP Sovereign is the customer-controlled edition of AICP, Veevra's AI Control Platform. Govern your models, knowledge and agents entirely in your own environment: private cloud, on-premises or fully air-gapped. Your models, your data, your control.
Requests can be routed, checked and traced through AICP. Hover any stage to see what happens there.
The enterprise AI challenge
Every department wants AI.
Coordination is the hard part.
Most organizations already run capable systems, and many teams now adopt AI in different places at once. The practical questions multiply quickly, and AICP exists to answer them in one place.
Which models, at what cost?
One registry and one bill for the models your teams use, with usage and quotas visible in one place.
Whose policy applies?
One set of guardrails, clearance and compliance rules, defined once and applied to AI use across teams.
Where does the data go?
You choose the providers and deployment model, so sensitive content stays within the boundaries you set.
Can a result be explained?
Requests routed through AICP are recorded with their context, so a result can be reviewed and explained later.
The AI Control Platform
Connect once.
Power everything.
AICP is a customer-controlled AI platform. Connect your applications and AI capabilities to it; from there, AI can be governed, checked and traced. You keep your applications.
AI Control Platform · AI Service Bus
Each connection runs through the governed, audited path.
How AICP works
One control plane.
Every capability, governed.
Inside AICP, the AI Service Bus is the single entry point. Routing, policies, knowledge, models, agents, OCR and audit are not separate tools you integrate; they are building blocks of one platform. Hover any capability to see what it gives every experience built on it.
AI governance
Policy enforced
on requests routed through AICP.
Checks run before results are returned
Guardrails run on input and output: prompt-injection screening, PII detection and redaction, content filtering, clearance and data-residency checks. A human approval gate steps in wherever your policy requires judgment, and the configured checks complete before a result is returned.
Knowledge intelligence
Ask a question. Get an answer with its sources.
AICP retrieves from your own documents and data and answers with citations back to the passages it drew from, so a reviewer can check where an answer came from before relying on it.
- Document Intelligence & OCR
- Recognition, classification, extraction and redaction turn documents into structured, governed knowledge.
- Semantic Search & RAG
- Hybrid search and retrieval-augmented generation draw responses from your own sources.
- Knowledge Graph
- Entities, facts and relationships connected across systems, so answers carry context.
- Cited by design
- Answers carry citations to the document and passage they drew from.
AI agents
One task, run as a governed trace.
Agents plan and execute multi-step tasks behind the policy checks and approval gates you configure. An illustrative walkthrough of one supported run:
- Task"Prepare a compliance summary for this tender." The agent plans the steps.
- Permitted tool actionsIt searches your knowledge collections and reads the tender through governed connectors, only what policy allows.
- Approval where requiredBefore anything high-impact, the run pauses for human sign-off, with the context needed to decide.
- Recorded resultThe summary is delivered, and each step of the run, tools, sources, approvals, is written to the audit trail.
AI automation
From trigger to action,
governed end to end.
AICP turns any event into an intelligent, automated process, with a person in the loop only where judgment is required.
- TriggerAn email, form, upload, API call or schedule.
- AI UnderstandingAICP reads, extracts and classifies the content.
- Policy CheckGovernance, compliance and data-residency rules apply.
- Workflow DecisionThe process routes, branches and assigns by your rules.
- if neededHuman ApprovalA person steps in only when judgment is required.
- ActionA task runs, a system updates, a document or alert goes out.
- Audit TrailEach step is written to the audit trail.
Enterprise integrations
One governed API
between your systems and AI.
AICP sits between your systems and the AI that serves them, brokering the connections, so applications call one governed API instead of wiring to models directly.
Plus ServiceNow, Active Directory, LDAP, SQL databases and REST / custom APIs, through standard, supported interfaces.
Add a provider or a system once. Every governed experience can use it immediately.
Deployment models
Run AICP
where your data lives.
On-Premises
Your data center, your infrastructure and models.
Air-Gapped
Fully disconnected, on-host models, no external calls.
Public or Private Cloud
Your own public or private cloud, with complete data-residency control.
Hybrid
On-prem for sensitive work, cloud or national AI where it fits.
AICP is customer-controlled: you deploy and run it in your own environment, it is not a SaaS. For a Veevra-operated AI platform delivered as SaaS, see AICP Enterprise.
Business use cases
One platform.
Every intelligent process.
V-Workspaces can be empowered by AICP for customer-controlled deployments or by AICP Enterprise for SaaS.
- Documents & Correspondence
- Document Intelligence · OCR & Extraction · Classification · Entity & Metadata · Redaction · Correspondence
- Contracts, RFPs & Compliance
- Contract Intelligence · RFP & Tender · Compliance Assessment · Risk & Obligations · Governance Review
- Knowledge & Search
- V-GPT Enterprise Chat · RAG Assistant · Semantic Search · Enterprise Search · Knowledge Graph · Business Facts
- Operations & People
- Executive Intelligence · Research · Meeting Minutes · HR & CV Review · Procurement · Project Health
- Automation
- Agent Execution · Workflow Execution · Human Validation · Translation · Summarization
- Trust & Platform
- Policy Evaluation · Explainability / Trace · Audit Trail · Cost & Usage · Model Routing · Multi-Model Compare
See it on your systems
Get a tailored walkthrough
Leave your details and our team will show you Veevra working in your environment, with no commitment.
See governed AI, end to end.
From a real document to an explained, audited answer. Book a guided demo, or experience AICP as a cinematic walkthrough.
Frequently asked questions
What is AICP Sovereign?
AICP Sovereign is the customer-controlled edition of AICP, Veevra's AI Control Platform. You deploy and operate it entirely in your own environment to govern your models, knowledge and agents.
Can AICP run on-premises or air-gapped?
Yes. AICP Sovereign is designed to run in your private cloud, on-premises or fully air-gapped, as well as in hybrid setups.
Does AICP Sovereign support sovereign AI?
Yes. With AICP Sovereign your models, data and control stay inside your own environment, so nothing has to leave your infrastructure.
Who operates AICP Sovereign?
You do. Unlike AICP Enterprise, which Veevra operates as SaaS, AICP Sovereign is run by your own team, with Veevra available for consulting, managed services and support.
Does AICP Sovereign support AI governance and agents?
Yes. It gives you one governed control plane for the models, knowledge and agents your organization uses, with your policies applied by default.