# Vendor Risk Questions

Questions to ask before enabling AI features.

Track: AI Vendor Evaluation

## What is it?

Vendor risk questions are the practical questions you ask before turning on an AI feature from a software provider. They help you understand data, accuracy, controls, transparency, ownership, and support.

## Why it matters

A vendor AI feature can change how your team works and how your data moves. Good questions help you avoid surprises. They also help legal, security, privacy, procurement, and business teams review the same facts instead of arguing from assumptions.

## How it works

Group the questions into simple areas: what the feature does, what data it uses, how output is checked, what controls admins have, what evidence the vendor provides, and what happens when the feature fails or changes.

## Analogy

Think of vendor AI review like renting office space. You do not only ask if the building looks good. You ask about locks, access, exits, maintenance, liability, and who to call when something breaks.

## Example usage

Before enabling AI in a CRM, ask whether customer data is used, whether admins can control the feature, whether prompts are logged, whether outputs are explainable, and whether the vendor has a clear support path for errors.

## How to use this

Keep a short standard question set for vendor AI. Use it before enabling features, renewing contracts, or expanding licenses. The same question set makes reviews faster over time.

## Common mistake

The common mistake is asking only, is it secure? That question is too broad. Ask specific questions that produce specific answers.

## Question to ask

- **Function**: What exact task does the AI feature perform?
- **Data**: What data does it process and where?
- **Control**: What can administrators limit, disable, or audit?
- **Change**: How will we know if the AI feature changes?

## Quick quiz

Why use standard vendor AI questions?

## Flashcard

**Question:** What are vendor risk questions?

**Answer:** Vendor risk questions are a standard set of practical questions used before enabling or expanding vendor AI features.