Data location
Determine whether prompts, transcripts, documents, embeddings, model outputs, and analytics records remain in the cloud or inside the enterprise.
ENTERPRISE AI DEPLOYMENT
Deploy enterprise AI on your terms.
Run CBOT through SaaS, private cloud, hybrid, or fully on-premise architectures, according to your data, security, infrastructure, regulatory, and operating requirements.
Choose where your models run, where your enterprise data remains, and how your AI environment is managed, without giving up the capabilities of the CBOT platform. One enterprise AI platform. Your operating model.
WHY DEPLOYMENT CHOICE MATTERS
Where AI runs changes what the enterprise controls.
Deployment architecture affects more than infrastructure. It determines where sensitive data is processed, which models can be used, how services are operated, and how the platform connects to internal systems. Different organizations, and even different workloads, may require different levels of control.
Infrastructure control
Choose whether infrastructure is managed as a service, operated in a private cloud, or controlled entirely by the organization.
Model location
Use managed model services, private endpoints, customer-hosted models, CBOT models, or supported open-source models by workload.
Regulatory & security
Design the architecture around internal security policies, data-residency requirements, sector regulations, and approved environments.
Deployment is not only where the software runs. It defines the boundary of control.
ONE PLATFORM, FOUR DEPLOYMENT MODELS
Choose the operating model that fits your enterprise.
The same CBOT platform capabilities remain consistent while the infrastructure boundary changes. Select a model to see where the platform, models, and data operate.
01
SaaS
Use CBOT through a managed cloud service for workloads approved for external operation. Reduce infrastructure preparation and accelerate the path from design to production. Best for speed to production, cloud-approved workloads, and a managed operating model. Enterprise AI without managing the complete underlying infrastructure.
AIFlow01 / 04
COMPLETE PRIVATE AI
On-premise means the complete stack, not only the application layer.
Some platforms run their application layer privately while still depending on external services for models, speech, retrieval, or vector storage. CBOT can bring the complete enterprise AI stack into the customer-controlled environment. (CBOT models, supported open-source, and customer-hosted models run privately; commercial models remain available only through external or private endpoints.)
01Experience
VoiceWebMobileMessagingEmployee appsWorkflows
02Agent & application
AI agentsOrchestrationBusiness rulesHuman approvalIntegrations
03Intelligence
CBOT LLMPrivate & open-source modelsProprietary NLP/NLUSpeech-to-TextText-to-Speech
04Knowledge
Enterprise RAGDocuments & dataVector storagePrivate processing
05Runtime & infrastructure
Model inferenceCPU & GPUContainersKubernetes / OpenShiftMonitoring & logsIndependent scaling
No mandatory external AI service between your data and your outcome.
One enterprise can use more than one deployment model.
Deployment doesn't always need to be a single organization-wide decision. Different agents, models, channels, and workflows can operate in different environments according to their own requirements, each routed only to an approved environment.
- Public information agentMay use approved cloud infrastructure for non-sensitive information.
- Customer account processKeeps customer data, enterprise knowledge, and workflow execution inside the private environment.
- Complex reasoning taskMay use an approved commercial model through a controlled endpoint.
- Sensitive document processingUses a customer-hosted or on-premise model without transferring documents outside the boundary.
- Real-time voice agentRuns speech and inference where latency, privacy, and capacity requirements fit best.
Choose deployment at the workload level, not only at the platform level.
Define who operates every layer.
Deployment selection also determines how infrastructure, platform services, model capacity, monitoring, security, and updates are operated. CBOT works with enterprise teams to define the appropriate responsibility model. Select a deployment model to see how responsibilities shift.
CBOT
Shared
Customer
Infrastructure provisioning
CPU & GPU capacity
Network configuration
Container environment
Platform deployment
Model deployment
Monitoring & logging
Backup & recovery
Security policies
Access management
Platform updates
Capacity planning
Representative responsibility model. The final operating model is defined per agreement during solution design.
The operating model is designed as carefully as the technology architecture.
Which deployment model fits your requirements?
Select the priorities that matter most to your organization. The result is a suggested starting point, not a final architecture decision.
SUGGESTED STARTING POINT
Select one or more priorities to see a suggested starting point.
Final deployment architecture should be defined through technical and security assessment.
Request a deployment assessmentKeep every workload inside its approved boundary.
CBOT deployment architectures can be designed around enterprise security, access, retention, network, and data-processing policies, controlling where data enters, where it is processed, and where it is stored.
- Private network connectivity
- Role-based access
- Customer-controlled identity systems
- Encrypted storage & transmission
- Configurable data retention
- Data masking & anonymization
- Audit trails
- Approved model access
- Controlled external endpoints
- Customer-managed secrets & credentials
- Private logging & analytics
- On-premise data processing
Control where data enters, where it is processed, and where it is stored.
Frequently asked questions
What deployment options does CBOT support?
CBOT supports SaaS, private-cloud, hybrid, and fully on-premise deployment models. The appropriate architecture depends on your security, data, infrastructure, performance, and operational requirements.
Can the complete CBOT platform run on-premise?
Yes. The platform, AI agent runtime, supported model inference, proprietary NLP, Speech-to-Text, Text-to-Speech, RAG, vector storage, analytics, and integrations can operate within the customer's environment.
Can CBOT operate without external commercial model APIs?
Yes. Organizations can use CBOT models, supported open-source models, and customer-hosted models without a mandatory dependency on an external commercial model API.
Can cloud and on-premise models be used together?
Yes. Hybrid architectures can combine approved cloud services with private, customer-hosted, and on-premise models, routing each workload by its security, performance, data, and model requirements.
What is the difference between private cloud and on-premise?
Private cloud typically operates within an isolated cloud environment. On-premise operates inside infrastructure controlled directly by the organization, such as its own data center. The exact responsibility model depends on the selected architecture.
Can enterprise data remain entirely on-premise?
Yes. In a fully on-premise architecture, prompts, conversations, voice streams, documents, knowledge, embeddings, model outputs, and analytics records can remain within the approved enterprise environment.
Does on-premise deployment require customer GPU infrastructure?
GPU requirements depend on the selected models and workloads. Language model and selected speech workloads may require customer-controlled GPU, while other services can run on CPU. Detailed sizing is completed during solution design.
Does CBOT support Kubernetes and OpenShift?
Yes. CBOT supports containerized deployment using Docker, Kubernetes, and Red Hat OpenShift environments.
Can we move from one deployment model to another?
Deployment architecture can evolve as requirements change. The migration approach depends on the existing environment, models, integrations, data, security requirements, and operational design.
How should an enterprise select a deployment model?
Consider data sensitivity, regulation, infrastructure strategy, model requirements, latency, concurrency, GPU availability, operational responsibility, and expected growth. CBOT can support a technical and security assessment to define the appropriate architecture.
YOUR AI. YOUR OPERATING MODEL.
Choose where your enterprise AI runs, and keep control of what matters.
Talk to CBOT about your security, data-residency, infrastructure, model, performance, and operational requirements. We can help you design the right SaaS, private-cloud, hybrid, or fully on-premise architecture.