On-premise, our own and open-weight models, and AI consulting
Dialogist on your servers, inside your network
Dialogist Enterprise is the full platform deployed inside your own environment. It runs on our own and open-weight models on your own servers, so no caller audio or text goes to a third-party AI provider, and nothing leaves your network that you decide should not.
With the deployment come AI and ML consultants, who design the conversation model with your team, build the evaluation sets and stay through the rollout.
Capabilities
On-premise deployment
The complete platform (control layer, voice engine and knowledge engine) running inside your infrastructure, behind your network boundary and under your change-management process. Speech recognition, language and voice models run on your servers too, and an air-gapped installation is supported.
Model fine-tuning on your data
PlannedSpeech recognition fine-tuned on your own recordings and call transcripts, so the agent understands your terminology and the way your customers speak. It will run in Cadence, the workbench we train Oasis Speech in, inside your environment.
AI and ML consulting
Conversation design, evaluation sets, prompt and model strategy, and deployment planning, worked through with your team rather than handed over as documentation. We define the metrics that tell you whether it is working.
Fully isolated instances
A dedicated installation for your organisation, operationally separated from every other customer, for organisations whose requirements do not permit shared infrastructure. In the standard cloud each organisation's data is kept separate.
Regional data residency
The Dialogist cloud is hosted in the EU. For markets that require data to stay elsewhere, an isolated regional cluster with its own voice and knowledge engines and storage is set up on request.
Large call volumes
Hundreds of simultaneous calls. Each Server Unit carries 100 simultaneous calls and has been load-tested to 176% of that; we size the number of units, the models and the architecture per project, as an enterprise engagement with its own pricing.
Enterprise SSO with SAML
PlannedSingle sign-on against your own identity provider.
Automatic dedicated environment per organisation
PlannedA fully isolated deployment provisioned automatically rather than as a services engagement.
Questions
What does on-premise include?
The control layer, the real-time voice engine and the knowledge engine, all deployed in your environment. You keep operational control; we continue to maintain and update the platform.
Can the models be trained on our own data?
The agent learns your business from your documents through the Knowledge Hub. Fine-tuning Oasis Speech on your own recordings, inside your environment, is on the roadmap: Oasis Speech, our own recognition service, is trained in Cadence, our training workbench, and that data would not leave your infrastructure.
How does pricing work at this level?
It is scoped per engagement rather than sold as a published list price, because it depends on deployment shape, volume and the consulting involved. Contact us.
Who does the AI and ML consulting?
Dialogist's own data science and conversation design team, working with yours. We do not resell a third-party AI service and we do not hand the work over to a partner: the models are our own and open-weight, run on our servers or yours, and no caller audio or text goes to a third-party AI provider.
Next step
Dialogist on your servers, inside your network
Join the waiting list
We're letting teams in gradually. We'll email you when your spot opens.