Inquio User Manual

Inquio reads the conversations your virtual assistant has already had and tells you what is working, what is broken, and what to fix first. You upload a transcript file. Inquio analyzes every conversation in it and returns a graded report card, a ranked list of problems, and step-by-step recommendations.

There is no dashboard to build and no query language to learn.

Who it is for

Inquio is written for the people responsible for how an assistant behaves: product owners of chatbot and voicebot solutions, customer care managers, and CX managers. If you own the answer to "why are users unhappy with our bot", this is your tool.

What makes it different

Most analytics tools count things. Inquio judges them.

It reads a conversation the way a reviewer would and decides whether the user actually got what they came for. That is why the output is a graded report card rather than a chart. The hard part is not measuring volume, it is knowing which conversations went wrong and why.

How it fits together

Your organization contains one or more bots. Each bot collects report cards.

A report card is a frozen snapshot of one upload, covering one dataset and one time frame. Every upload produces a new one, so a bot's cards form a record of how it changed over time.

Live data works across report cards. When you want to look at a specific day or a period that spans several report cards, switch to Live data and you get the same statistics, issues, and conversations for exactly the range you choose, without a report card attached.

From there, seven screens go progressively deeper:

ScreenAnswers
Bot Report CardHow is my bot doing overall?
Issues and RecommendationsWhat should I fix, and in what order?
ConversationsWhat actually happened in these conversations?
User utterancesHow do users phrase what they want?
Conversation metricsHow much traffic, when, and how efficiently handled?
DefenderIs my bot behaving safely, and is it under attack?
Sales RadarWhich customers should sales talk to?

What one upload gives you

  • An Inquio Score™ with four graded dimensions: Resolution Accuracy, Containment Quality, Dialog Fluency and Safety & Compliance
  • A comparison against the average for your industry
  • Issues ranked by severity, each with recommended steps and the predicted effect on your score
  • Safety incidents and attempted attacks
  • Sales leads and churn risks

The Starter plan is free for up to 1,000 conversations, so you can run a real dataset through it before you decide. See Plans and limits.

What Inquio does not do

Inquio analyzes transcripts. It does not build, host or connect to your assistant. It does not change anything in it, and it does not alert you in real time (Enterprise plans can send data through the API instead of a file). You act on what it finds in whatever platform your assistant runs on.

A note on input and languages

The interface is in English. Conversations can be in any language.

Input must be text, but the source does not have to be a chat window: transcribed voice calls work, and so do message threads, and emails between two people.

Deep language analysis (the actions and objects on the User utterances screen) is built for English and Czech. For other languages, scores, issues and topics work as usual; contact us to enable deep analysis.

Where to go next