Spaces

Is your program office ready for an AI agent? A short quiz

This quiz asks eight yes or no questions about whether your program office is ready for an AI agent. It checks written answers, hand-over names, limits and a safe route for sensitive messages.

Runs in your browser. Nothing you enter is sent anywhere or saved.

Before you start

An AI agent works well in a program office when a few things are in place. Staff must agree what the agent may say, there must be a person to take over, and the guidance it reads must be accurate. Without those, an agent can do more harm than good, because the public treats its words as the office own.

Answer each question honestly with yes or no. There is no pass mark to worry about. The result shows what to prepare before you begin, and which steps are already done. Most offices that score low find they can reach a good level in a week or two of careful writing.

Do you have a current written summary of each program you run?
Do you have a list of forms with where to find the latest version?
Do you know which topics an agent must never answer, such as eligibility or health?
Do you have a named person or team to take over each type of sensitive chat?
Do you have a tested route for messages that suggest someone is in danger?
Do you know who will read and correct the agent replies each week?
Do you have a plan for keeping your information up to date when rules change?
Do you have a secure route for people to send documents and personal details?

Answer all 8 to see your result. 0 so far.

Where offices usually find gaps

The most common gap is not technical. It is that the information lives in staff heads, not on paper. Writing down hours, forms and the usual next steps is hard but valuable, and it helps new staff just as much as an agent.

The second gap is the hand-over route. Offices often agree that sensitive messages go to a person but have not said which person, or what happens at night. Settle that before launch and test it with a pretend message.

  • Write down what is currently in staff heads.
  • Name an owner and a backup for each hand-over type.
  • Test the urgent route end to end.

FAQ

Do we need to be perfect before starting?

No. You need accurate basics, clear limits and a working hand-over route. Start small with routine questions, review the chats closely, and improve the guidance as you learn what people ask.

Is an AI agent suitable for a sensitive public program?

It can be, for the routine part, if decisions, health, safeguarding and personal records always go to a person. The agent supports staff with simple questions, and does not replace their judgement.

What should we do first if we score low?

Write one plain page for your biggest program, name who handles sensitive chats and list what the agent must not answer. That small set of steps lifts most offices to a safe starting point.

More free tools for education, health and social program offices

Read the guide: Answer routine program questions in chat, and keep every decision with your staff.