Demo · AI Agent
AI assistant demonstration
An AI agent accepts a natural-language request and produces a response using a language model. In a business setting, agents are useful for high-volume, text-heavy work: answering questions from an approved knowledge base, summarising threads, classifying requests, and drafting first responses for human review.
AI assistant — demonstration
Ask a question to see how the assistant responds. Suggested prompts:
What this demonstration does
- Uses the platform's real server-side endpoint, so a model must actually be configured to answer.
- Streams the response as it is produced.
- Shows an explicit unavailable state when the deployment has no model key — it does not fake replies.
- Keeps provider keys on the server; the browser never receives credentials.
What it does not do
- No browsing of the live web, no autonomous tool execution, no external actions.
- No persistent memory in this demo — no session id is sent, so conversations are not saved by the endpoint.
- No guarantees of accuracy. Output must be reviewed, especially for financial, legal, medical, employment, or security decisions.
- No voice, image, or file input in this demo.
Building with AI agents
A production agent usually needs grounding in your own documents, explicit scope limits, a review step for consequential outcomes, and cost controls. If you want to explore that for your business, send an enquiry describing the workflow, the inputs, and who reviews the output. You can also read what an AI agent is and when to use one.