Onplana Runs on Two AI Providers: Claude and Azure OpenAI
Onplana's AI runs on both Claude (Anthropic) and Azure OpenAI. Onplana manages the two providers for resilience and model fit, with automatic failover.
Most AI-native tools bet on a single model vendor. Onplana does not. Every AI feature in Onplana (risk detection, plan generation, the streaming chat panel, natural-language task parsing) runs on a dual-provider architecture: Claude (Anthropic) and the GPT-4 family via Azure OpenAI, side by side. Onplana manages which provider serves each request. There is nothing for you to configure.
This is a platform decision we make on your behalf, not a switch in your settings, and it buys you three things a single-vendor architecture cannot.
Why two providers
1. Resilience through failover
Model APIs have outages. When a tool depends on one provider, an incident on that provider means the AI features stop. Because Onplana runs both Claude and Azure OpenAI, a degradation or outage on one fails over to the other automatically, at the platform level. The AI keeps working while the affected provider recovers.
2. The right model for the shape of the work
Claude and the GPT-4 family are both strong, but not identical. Some workloads (long-context synthesis, carefully hedged risk narrative) suit one; others (tightly structured output, concise executive summaries) suit the other. Running both lets Onplana route a given kind of work to the model that handles it best, without asking you to think about any of it.
3. No single-vendor lock-in
AI vendors move fast and their terms change. A tool welded to one model inherits that vendor's pricing, availability, and policy shifts wholesale. A dual-provider architecture gives Onplana room to adapt the routing as the model landscape changes, so the capability you rely on does not hinge on one company's roadmap.
How it works, and who runs it
The important thing to be clear about: provider selection is managed by Onplana, not by your workspace admins. There is no provider toggle in your settings, and there is no "bring your own Azure OpenAI" step for you to set up. Onplana operates both providers and routes between them.
A few specifics on the security posture, since enterprise buyers ask:
- Credentials. Onplana's provider API keys (Anthropic and Azure OpenAI) live in Azure Key Vault, rotated by ops out of band. They never appear in the product UI.
- Data handling. Onplana sends prompts to Anthropic and Azure OpenAI under their enterprise API terms, which do not train on customer inputs. Your project data stays in your organization's database; the AI reads it at request time, and the providers do not retain it beyond the request.
- Nothing for you to run. Because Onplana operates the providers, there is no Azure resource for your team to stand up, no keys for you to supply, and no deployment step on your side.
What this means for you
Nothing to switch on. The dual-provider architecture is already how Onplana's AI runs, included in your plan rather than sold as a separate add-on. You get reliability when a provider has a bad day, model fit across different kinds of project work, and an architecture that is not hostage to a single vendor, all managed for you.
Questions about how Onplana handles AI, data, and security? Contact us or read the AI in Onplana overview.
Ready to make the switch?
Start your free Onplana account and import your existing projects in minutes.