Inquio Score™
The Inquio Score™ is the quickest way to understand how your bot is performing overall. It turns the conversations in one report card into a single score out of 100, a grade, and four supporting dimensions that explain where the result comes from.
- How is my bot doing overall?
- Which area is pulling the score down?
- Did the bot improve compared with the previous period?
- How does this bot compare with others in my industry?
The score is not meant to replace the detailed screens. It is a starting point. It tells you where to look first.
What the score measures?
The Inquio Score™ measures how well your bot handled the conversations in the selected report card.
It combines four dimensions:
| Dimension | Weight | What it asks |
|---|---|---|
| Resolution Accuracy | 40 % | Did the bot give correct and complete answers? |
| Containment Quality | 25 % | Were conversations resolved by the bot, rather than only kept inside the bot? |
| Dialog Fluency | 20 % | Did the conversation flow naturally, without avoidable repetition, confusion, or dead ends? |
| Safety & Compliance | 15 % | Did the bot behave safely, follow policy expectations, and resist inappropriate or adversarial use? |
Because the dimensions have different weights, they do not affect the final score equally.
Resolution Accuracy has the largest effect. For example, a 10-point improvement in Resolution Accuracy moves the total score by 4 points. A 10-point improvement in Safety and Compliance moves it by 1.5 points.
That does not mean safety is less important. It means the overall score is weighted toward the areas that most often decide whether users get what they came for.
Resolution Accuracy
Resolution Accuracy measures whether the bot’s answers were factually correct and complete.
- answers the question the user actually asked
- gives information that matches your product, pricing, policy, or process data
- includes the steps the user needs to act
- resolves the request without making the user ask again in a different way
- gives wrong information
- gives only a partial answer
- redirects the user instead of answering
- answers a related question, but not the one that was asked
Resolution Accuracy is the largest part of the Inquio Score™, worth 40% of the total.
Containment Quality
Containment Quality measures whether conversations that stayed with the bot were genuinely resolved.This is different from a basic containment rate. A conversation is not successful just because it did not reach a human agent. If the bot keeps the user inside the conversation but fails to solve the request, containment may look good while the user experience is poor.
Containment Quality looks for signs that the user’s need was actually handled, not just that the conversation remained automated.
It is worth 25% of the total score.
Dialog Fluency
Containment Quality measures whether conversations that stayed with the bot were genuinely resolved.
This is different from a basic containment rate. A conversation is not successful just because it did not reach a human agent. If the bot keeps the user inside the conversation but fails to solve the request, containment may look good while the user experience is poor.
Containment Quality looks for signs that the user’s need was actually handled, not just that the conversation remained automated.
It is worth 20% of the total score.
Safety & Compliance
Safety and Compliance measures whether the bot behaved safely and followed policy expectations.
- shares information it should not share
- responds inappropriately to sensitive topics
- speaks negatively about competitors
- fails to handle vulnerable or delicate situations with care
- follows an adversarial user into unsafe or policy-breaking behavior
This dimension also reflects whether users attempted to attack or manipulate the bot, and whether those attempts succeeded.
Safety and Compliance is worth 15% of the total score. A low score here can be serious even when the other dimensions look healthy, because one unsafe answer can create reputational, legal, or compliance risk.
For detail, open Defender.
Reading the score and the grade
The score is shown as a number out of 100 and as a grade.
| Score | Grade | What it means |
|---|---|---|
| 90-100 | A | The bot is performing very well. Focus on small gaps and maintaining Quality. |
| 80-89 | B | The bot is mostly reliable, with recurring issues in one or two areas. |
| 70-79 | C | The bot is useful, but there are visible weaknesses that users will notice. |
| 60-69 | D | A substantial share of conversations have Quality problems. |
| 0-59 | F | Users cannot reliably depend on the bot to handle their requests. |
A high grade does not mean there is nothing to improve. It means the most important problems are likely narrower and easier to prioritize.
A low grade does not mean you need to rebuild the whole bot. It means you should open the weakest dimension, review the issues behind it, and work through the recommendations by impact.
What the score does not measure
The Inquio Score™ is a Quality score for the conversations in a report card. It does not measure everything about your assistant or your business.
- total traffic
- number of users
- cost savings
- customer lifetime value
- agent workload
- infrastructure performance
- whether a specific product or campaign is successful
Some of those questions are answered elsewhere in Inquio. For example, use Conversation metrics to understand traffic and workload, Sales Radar to find opportunities and churn risks, and Conversations or User utterances to inspect what users are actually asking.