Use AI within the boundaries your organization approves.
Portupine supports Anthropic, Google Gemini, OpenAI, OpenAI-compatible endpoints, and local Ollama models. Choose the provider, model, and data boundary for each deployment while keeping AI output advisory and subject to human review.
Choice
Choose the model boundary
Connect an approved provider, an OpenAI-compatible endpoint, or a local Ollama model.
Data boundary
Keep prompts in approved infrastructure
A local configuration can run AI workflows without sending data to an external model provider.
Human control
Advisory by design
AI drafts and summarizes; accountable people review and make the decision.
AI in hiring.
Automatic CV parsing and fit-scoring on application, interview question generation, and candidate ranking — always advisory, never auto-rejecting.
AI on your accounts.
Client summaries, drafted emails, next-action suggestions, and churn risk scoring across your CRM.
AI in reviews.
Draft summaries that synthesize self and peer feedback into editable themes for managers — the manager writes the review; the AI does the reading.
AI in conversations.
Summarize long chat threads on demand.
Honest by design.
If no provider is configured, AI features switch off gracefully. If a provider fails, you get a clean error — never a fabricated answer.
In the product
Advisory output, next to the data it came from
The client workspace pairs the account record with an AI summary and churn risk score — a second pair of eyes the account team reviews, not a decision made for them.
Per-tenant provider, model, and key configuration; secrets stored encrypted and masked in the UI.
Scope the platform around your organization.
Discuss the systems you want to consolidate, deployment boundaries, rollout priorities, and support requirements.