Platform Architecture
The Enterprise AI Platform brings agents, orchestration, connectors, governance, and deployment together under a shared foundation, structured in four layers:
- Application Layer: Generative AI interfaces, copilots, and workflow-specific custom apps.
- Orchestration Layer: Autonomous AI agents and collaborating multi-agent teams built on the AI Agent Factory foundation.
- Connector Layer: API connections linking agents to corporate databases, document hubs, ERP, and legacy systems.
- Governance & Operations Layer: Workspace permissions, model evaluation pipelines, audit logs, and drift monitoring.
Access Boundaries
Tenant isolation and role-based access are part of the platform's governance layer - see Security & Governancefor the full architectural detail. The questions below separate what's documented as platform capability from what needs confirming for your specific deployment.
How is tenant data isolated from other customers?
Role-based access control and per-tenant isolation are part of the platform's governance layer - see Security & Governance for the architectural detail. The exact isolation guarantee for your specific deployment model (SaaS multi-tenant vs. BYOC/on-premises single-tenant) should be confirmed during scoping rather than assumed uniform across all models.
Who can access model inputs and outputs - LuMay staff, or only our team?
This depends on deployment model and support arrangement, and is not stated as a fixed, universal answer in LuMay's published materials. Confirm the exact access model - including any support-access exceptions - as part of your platform evaluation.
Where is telemetry and model-usage data stored, and for how long?
Observability (tracing, metrics, alerting tuned to latency and cost) is part of the platform, instrumented during delivery. Retention period and storage location for telemetry data specifically should be confirmed for your engagement - this page does not state a fixed retention policy because none is published site-wide.
Deployment Dependencies
The platform is container-portable: the same images run across every deployment model, so the governance layer stays consistent regardless of where it runs. See Deployment Models for the full comparison.
- SaaS: LuMay-managed hosting - the fastest path to pilot and production.
- Your Cloud (BYOC): Deployed inside your own Azure, AWS, or GCP subscription. You choose the region and control the infrastructure boundary.
- Private Cloud / On-Premises: An annual license inside your own infrastructure, with security review as part of onboarding.
- Air-Gapped: For isolated environments with no external connectivity - custom architecture and an offline update process.
Operational Ownership
Who operates the platform day-to-day after go-live?
The deployment fabric, connectors, and governance layer are configured as part of the engagement, and runbooks are handed off for incident response, capacity planning, and routine operations. Whether LuMay continues an ongoing operating role, or your team takes over day-to-day operations, is a decision to confirm during scoping, not a platform default.
Who is responsible for model evaluation and drift monitoring once the platform is live?
Drift monitoring is part of the governance layer, but ownership of acting on what it surfaces - re-tuning, retraining, or escalating - should be assigned explicitly as part of your operating model, the same way infrastructure ownership is.
What happens if we need to change deployment model later (e.g. SaaS to BYOC)?
Not addressed as a standard migration path in LuMay's published materials. Raise this directly if it's a plausible future requirement for your organisation - for example, starting on SaaS for a pilot and moving to BYOC for a UK/EU data-residency requirement once the business case is proven.
Related: Enterprise AI Development for the engagement model that typically precedes standardising on this platform, and UK pricingfor the Enterprise Platform bundle's published starting price.
FAQ
Does the Enterprise AI Platform guarantee UK or EU data residency?
Not as a default. The platform supports BYOC deployment inside your own Azure or AWS subscription, private cloud, and on-premises/air-gapped options - these let you choose where data is processed and keep it within a region you control. End-to-end UK data residency is not asserted on this page. If residency is a hard requirement, confirm the specific hosting, processing, and related data locations for your deployment directly with LuMay before relying on any particular region.
Is the platform multi-tenant or single-tenant?
This depends on deployment model: SaaS is managed by LuMay; BYOC, private cloud, and on-premises deployments run inside infrastructure you control. Confirm the isolation model that applies to your chosen deployment during scoping.
Does LuMay hold any UK-specific certifications for this platform?
Not stated in LuMay's published materials. If a specific certification (e.g. Cyber Essentials, ISO 27001 scope covering a UK entity) is required for your procurement process, confirm directly rather than assuming it from this page.