A clear role. Clear boundaries.

Pre-launch draft. These policies describe proposed arrangements, not a service operating today. Product behaviour, commercial terms, company details and legal review remain pending. They do not take effect until launch.

This website is a scripted preview. The live service, its providers and its data arrangements still need to be finalised and tested.

The intended conversation journey

The visitor chooses voice or written chat after a clear AI introduction.
Voice asks for microphone permission. A visible transcript and stop control are part of the intended experience.
The assistant uses information the business has approved and asks for clarification when needed.
An approved action needs visitor confirmation and a successful response from the connected system before it can be called complete.
When a person is requested, and before collecting an enquiry, it explains which details are needed and asks permission to share them.
The business receives an English summary only after the agreed delivery system confirms success. Staff review the request and confirm the next step.

What this preview does today

The conversation examples are pre-written. They do not open your microphone, contact an AI provider or submit the fictional enquiry. The contact button opens your email application; you choose whether to send the email.

This is a WordPress website. Hosting and platform processing are described in the draft cookie policy and draft privacy policy; a pre-launch review is still needed.

What needs to be agreed before launch

The minimum enquiry fields, consent wording and authorised recipients.
Which session and usage records are necessary, where they are held and for how long.
The provider contracts, processing locations and handling of audio and text.
Deletion, access requests, security controls and the process for ending service.

The intended design avoids storing voice recordings. This is a design requirement to validate, not a claim about an already deployed Punjabi Labs assistant.

Useful assistance, with limits

AI can misunderstand speech or give an incorrect answer. Approved information, clear uncertainty and a route to staff are safeguards that must be tested; they cannot guarantee perfect answers.

The planned assistant provides business information and collects enquiries. It does not make professional decisions or invent availability, prices or outcomes. Emergency and regulated-advice workflows are outside the initial scope.

Verify every commitment

A booking request is not a confirmed booking. Confirmation requires a connected system that checks availability, saves the booking and returns success, or a person who confirms it. The current examples include direct answers, simulated booking success and a handover explicitly requested by the visitor. Demo references beginning DEMO are fictional and do not reserve anything.

Questions about your workflow

Email contact@punjabilabs.com to discuss the intended data journey. The data processing terms and privacy policy remain drafts pending review.