Skip to main content
OFFSTAGE

OFFSTAGE

About our AI

OFFSTAGE uses AI to help run events. Here, in plain words, is what it does, what it never does and what happens to your data.

  1. Agents propose
  2. Policy decides
  3. Humans approve
  4. Code executes

AI suggests, people decide

Agents draft plans, messages and changes. Anything that affects people, money or the event's reputation waits for a named person to approve it. Small internal records run on their own, and those can be undone for 10 minutes.

How much approval a change needs

  • Risk tier T0: Internal record, runs on its own
  • Risk tier T1: Runs on its own with a 10 minute undo
  • Risk tier T2: Needs approval from the lead
  • Risk tier T3: Needs two approvals

Every automated message says so

Messages written by the assistant carry a line saying it drafted them and which role approved them, like this:

Drafted by the OFFSTAGE assistant, approved by Event lead

It says “I don't know”

The helpdesk answers only from the event's own documents and shows where each answer came from. When it is not sure, it passes your question to a person instead of guessing.

Emergencies go to people

Medical, safety, fire and harassment reports skip the agents and alert the organisers at once. Agents only help draft updates.

Minimum data, deleted after the event

We collect only what the event needs and say why when we ask. Personal details are deleted 90 days after the event unless you asked for a certificate. Phone numbers and emails are encrypted, and the AI works with IDs and masked details, not your contact information.

Your data, your choice

From your attendee portal you can see what we store, download a copy, or ask us to delete it.

What we do not claim

OFFSTAGE is designed against the OWASP Top 10 for LLM applications and aligned with the principles of India's Digital Personal Data Protection Act. It is not certified by anyone. It works in English, Hindi and Hinglish; other languages are not supported yet.