Glossary

Every term Inquio uses, defined in one line, with a link to the page that explains it properly.

A

Abandoned conversation — A conversation the user left before their request was resolved, by closing the chat or hanging up. See Conversation metrics.

Action — In language analysis, what the user wants done, expressed as a verb in its base form: activate, cancel, display. See Topics, actions and objects.

Admin — The role that can change the plan, see invoices, and add or remove people. One per organization. See Organization and users.

Analyst — The role everyone else has. Full access to bots, data and findings; no access to billing or user management.

Assistant — Your chatbot or voicebot. Inquio also calls this a bot; the two mean the same thing.

Attack — A user deliberately trying to make your assistant misbehave, reveal information it should not, or damage your reputation. Recorded as successful or unsuccessful. See Defender.

B

Benchmark — The average score for the industry chosen when the bot was created. Your bar against theirs, on the report card.

Bot — See Assistant.

Bot misbehavior — A Defender category for incidents the assistant caused itself, rather than ones a user provoked.

Bot Report Card (BRC) — A frozen snapshot of one upload: score, dimensions, issues, benchmark and recommendations for one dataset and period. See Your first report card.

Bounced — A conversation where the chat opened and the user never replied.

C

Card — Short for Bot Report Card.

Channel — How an utterance was entered: VOICE, TEXT or BUTTON. One of the five required CSV columns. See Prepare your data.

Churn — A signal that a customer intends to leave: closing an account, cancelling a contract, moving to a competitor. See Sales Radar.

Containment quality — Whether conversations that stayed with the assistant were genuinely resolved, rather than merely not escalated. 25% of the Inquio Score. See Containment quality.

Containment Quality Issue (CQI) — A flag on a conversation that was kept by the assistant without being resolved.

Conversation — One complete exchange between a user and your assistant, identified by its session_id.

custom_meta — An optional CSV column carrying any parameter or flag for a single utterance — a thumbs up or down, a client ID — displayed alongside the conversation. See CSV schema.

D

Defender — The screen covering safety: incidents the assistant caused, and attacks users attempted. See Defender.

Dialog Quality Issue (DQI) — A flag on a conversation where the exchange itself broke down — repetition, misunderstanding, failure to recover.

Dialog fluency — Whether the conversation flowed naturally. 20% of the Inquio Score. See Dialog fluency.

Duplicate — A conversation whose session_id is already in Inquio. Skipped on upload and not charged. See Upload more data.

E

Empty conversation — A conversation where the user never sent a message at all. A placement and greeting problem rather than a dialog one.

Enterprise — The negotiated plan: everything in Growth plus SSO, admin controls, API access and a support SLA.

Expanded object — The object as the user actually phrased it, with its modifiers: savings account, my new debit card.

F

FCR — First Call Resolution. The share of conversations on a topic where the request was resolved. Shown per topic on Conversation metrics.

Flag — What Inquio attaches to a conversation to describe what happened in it: Kudos, Neutral, or one of the issue types. A conversation can carry several at once. See Flags and issue types.

Frequent topics — The most common topics in a dataset, with the mix of outcomes in each.

G

Grade — The letter beside a score: A to F, on the overall Inquio Score and on each dimension.

Growth — The pay-as-you-go plan, billed per conversation analyzed.

I

Impact — The predicted change to your Inquio Score if you act on a recommendation, or the cost of an incident.

Incident — A safety or policy failure found in a conversation. See Defender.

Industry — Chosen when a bot is created and used to pick its benchmark. Cannot be changed afterwards.

Inquio Score™ — The overall score out of 100, built from four weighted dimensions. See Inquio Score™.

Issue — A recurring problem Inquio found across conversations, ranked by severity and carrying a recommendation. See Issues and recommendations.

K

Kudos — A flag on a conversation where the user expressed satisfaction, thanks or appreciation.

L

Leads — Conversations showing buying interest, worth passing to sales. Called Opportunities on older report cards. See Sales Radar.

Live data — A free time-range view across everything uploaded, independent of report card boundaries. Use it for a single day, or a period spanning two cards. Not available on Starter. See Live data and report cards.

N

Neutral — A flag on a conversation that went neither notably well nor notably badly.

O

Object — In language analysis, what the user's request is about, as a single noun in base form: card, password, account.

Opportunities — The former name for Leads.

Organization — The container holding your bots, report cards, data, subscription and users. Everyone in it sees all of it. See Organization and users.

P

Processing failed — The file was accepted and analysis began, but did not complete. Inquio's team is notified automatically. See When processing fails.

R

Recommendation — A specific fix written against what was found in your data, attached to an issue.

Report card — See Bot Report Card.

Resolution accuracy — Whether answers were factually correct and complete. 40% of the Inquio Score, and the heaviest lever on it. See Resolution accuracy.

Resolution Accuracy Issue (RAI) — A flag on a conversation where the assistant's answer was wrong or incomplete.

RT — Response time: the average seconds taken to resolve a request within one conversation, shown per topic.

S

Safety and compliance — Adherence to policy and resistance to adversarial users. 15% of the Inquio Score. See Safety and compliance.

Sales Radar — The screen covering revenue signals: Leads and Churn. See Sales Radar.

Session ID (session_id, SID) — The unique identifier of a conversation. Must be unique across every file you ever upload, not just within one.

Severity — How much attention an issue or incident deserves: Critical, High, Medium or Low. Reflects both volume and seriousness. See Severity levels.

Side — Who is speaking in a row of the CSV: USER or BOT.

Starter — The free plan. One bot, 1,000 conversations, one report card every 30 days, and no access to the conversations behind the findings.

T

Timestamp — When an utterance happened, to millisecond precision. One of the five required CSV columns.

Topic — A cluster of conversations about the same subject, generated by Inquio rather than defined by you. Broader than an action–object pair.

Trend — The arrow on a score, comparing this card against the immediately preceding interval of the same length — not against the previous card. See Trends and benchmarks.

Truncation — What happens when an upload exceeds your plan's limits: Inquio keeps the oldest data that fits and discards the rest, noting it on the card.

U

Unclassified — A flag on a conversation Inquio could not place in any other category.

User turns — The average number of user messages per conversation.

Utterance — A single message, from either side. One row of the CSV.

User utterances — The screen for analysing how users phrase their requests, by action and object. See User utterances.

V

Validation failed — The file was rejected before any analysis started. You can fix it and upload again immediately, at no cost to your allowance. See When processing fails.

Verbosity — The ratio of words written by the user to words written by the assistant. 1:9 means nine assistant words per user word.

W

Workload — Conversation volume by day of week and hour of day. See Conversation metrics.