Start with tenant isolation, not the marketing term "private"
"Private cloud" can mean very different things between vendors: a dedicated virtual machine still running the vendor’s shared multi-tenant database, a genuinely single-tenant deployment, or (rarer) a platform architected so each tenant’s data is isolated at the database schema level regardless of deployment mode. Ask specifically: is tenant data separated by a schema, or by an application-layer tenant ID filter on shared tables? Voz360 uses schema-per-tenant isolation — each tenant gets a dedicated database schema generated from one factory definition — in both managed SaaS and private cloud, so the isolation model does not weaken when you move to your own infrastructure.
Ask who owns the encryption keys
Encryption at rest is table stakes; the real question is who controls the keys. A platform where the vendor holds both the data and the only copy of the encryption key has not meaningfully changed your risk profile just because the servers sit in your data center. Look for a two-layer key design: a Data Encryption Key (DEK) unique to each tenant, itself wrapped by a Key Encryption Key (KEK) that can be managed separately from the data. Voz360 encrypts recording, voicemail, and SIP-secret content with AES-256-GCM under a per-tenant DEK, and every DEK is wrapped by a platform KEK using AES-KW (RFC 3394) — so no tenant’s key material is shared with another tenant’s.
Ask how the audit log proves tampering didn’t happen
Most vendors show you an audit log; fewer can show you why it can be trusted. An ordinary audit table is just rows in a database — anyone with sufficient access can edit or delete a row and leave no trace. Ask whether audit rows are cryptographically chained to each other, so that altering historical data would be mathematically detectable, not just procedurally forbidden. Voz360’s audit log SHA-256 hash-chains each row to the one before it, written under a row lock, so a verification pass can re-hash the chain and flag exactly where it breaks.
Ask what operational handoff actually looks like
Private cloud shifts genuine operating responsibility to your team: provisioning, scaling, patching, and monitoring. Before signing, get specific about what "you operate it" means in practice — sizing guidance, upgrade cadence, support escalation paths, and what remains the vendor’s responsibility versus yours. A platform built for pluggable deployment from the start should be able to describe this handoff concretely; a platform retrofitted for "on-prem" as a one-off enterprise deal usually cannot.
Confirm the architecture doesn’t fork between deployment modes
The single most useful confirmation is whether private cloud and managed SaaS run the same codebase. If they diverge — a "lite" on-prem version versus the full cloud product — you are evaluating two different platforms wearing one name, and feature parity claims should be treated skeptically. Ask for this in writing.
Can the vendor tell you — in one sentence — which of their AI capabilities are rule-based, which are generative, and which are still roadmap?