---
title: "Measure and improve AI"
description: "<p>Review resolution, involvement, handoff, quality, and knowledge-gap evidence.</p>\n<h2>How it works</h2>\n<p>Review resolution, involvement, handoff, quality, and knowledge-gap evidence."
---

> Documentation Index
> Fetch the complete documentation index at: https://tidyanswer.helpcent.re/llms.txt
> Use this file to discover all available pages before exploring further.

# Measure and improve AI

<p>Review resolution, involvement, handoff, quality, and knowledge-gap evidence.</p>
<h2>How it works</h2>
<p>Review resolution, involvement, handoff, quality, and knowledge-gap evidence. This article explains how the capability behaves in TidyAnswer, where to configure it, and what to verify before using it with customers.</p>
<h2>Set it up</h2>
<ol>
<li>Choose the inboxes and audiences where Tidy AI may participate.</li>
<li>Set approved knowledge, actions, boundaries, and handoff destinations.</li>
<li>Test confident answers, uncertain questions, and restricted requests before activation.</li>
</ol>
<h2>What to expect</h2>
<ul>
<li>The customer-facing behavior is predictable and reviewable.</li>
<li>Teammates can see ownership, state, and policy decisions in context.</li>
<li>Failures and handoffs remain visible instead of silently changing the journey.</li>
</ul>
<blockquote>
<p>[!TIP] Test changes with a controlled audience before applying them to every customer.</p>
</blockquote>

Source: https://tidyanswer.helpcent.re/ai-front-desk/measure-ai/index.mdx
