Call quality & conversation intelligence glossary

Short, accurate definitions of the key terms in call quality analysis, conversation intelligence, sales coaching and evidence-based interview analysis.

Conversation intelligence

Conversation intelligence is software that records sales or service calls and meetings, transcribes them, and analyzes the text and speech to surface insights such as topics discussed, talk ratios, questions asked, and next steps. It turns unstructured conversations into structured, searchable data for coaching and review.

Talk-to-listen ratio

Talk-to-listen ratio is the proportion of a call during which the salesperson speaks versus listens to the prospect, usually shown as a percentage split. It is derived from speaker-separated transcripts. Discovery calls generally favor more listening, while demos may involve more talking; the ideal balance depends on the call stage.

Talk time

Talk time is the total amount of time a specific participant speaks during a call, measured from a speaker-separated transcript. It is often expressed in minutes or as a percentage of total call duration. Talk time per participant is the basis for metrics like talk-to-listen ratio and longest monologue.

Longest monologue

Longest monologue is the duration of the single longest uninterrupted stretch of speech by one participant during a call. It is measured from the transcript by finding the largest continuous block before another speaker talks. A very long monologue can signal a pitch that talks at the prospect instead of engaging them in dialogue.

Filler words rate

Filler words rate is how often a speaker uses non-meaningful words and sounds such as um, uh, like, or you know, usually counted per minute or as a share of total words. It is calculated from the transcript. A high rate can make a speaker sound less confident or prepared, making it a common coaching target.

Speaker diarization

Speaker diarization is the process of partitioning an audio recording by who is speaking, answering the question who spoke when. It separates a single audio stream into segments attributed to different speakers, even without knowing their identities. Diarization is what lets a transcript label each line by speaker and enables per-speaker talk-time metrics.

Call transcription

Call transcription is the conversion of spoken audio from a call into written text, typically using automatic speech recognition. The output is a time-stamped transcript that can be read, searched, and analyzed. Transcription is the foundation for conversation intelligence: scoring, keyword search, and summaries all run on the transcript text.

Sentiment analysis

Sentiment analysis is the automated classification of text or speech as positive, negative, or neutral in tone. Applied to call transcripts, it estimates the emotional valence of what was said over the course of a conversation. It is an approximate signal based on language, not a definitive read of how a person actually felt.

AI notetaker

An AI notetaker is a tool that joins or processes a meeting, transcribes it, and automatically produces notes such as a summary, action items, and key topics. It removes the need to take notes manually during the call. The notes are generated from the transcript, so their accuracy depends on transcription quality.

Sales call QA

Call quality assurance (call QA)

Call quality assurance is the process of reviewing recorded calls against a defined set of criteria to measure how well agents follow a script, process, or standard. Traditionally a manager listened to a small sample; with AI, every call can be transcribed and scored against the same checklist consistently.

QA scorecard (rubric)

A QA scorecard is a structured list of weighted criteria used to evaluate a call, such as greeting, needs discovery, objection handling, and closing. Each criterion is scored, and the scores combine into an overall result. A shared scorecard makes evaluations consistent across reviewers and comparable across agents.

Sales call scoring

Sales call scoring is assigning a numeric or graded result to a sales call based on how it performed against defined criteria. Scores can be produced by a human reviewer or by AI reading the transcript. The aim is to quantify call quality so teams can track trends, compare reps, and target coaching.

Objection handling

Objection handling is how a salesperson responds when a prospect raises a concern or reason not to buy, such as price, timing, or competition. Effective handling acknowledges the concern, clarifies it, and addresses it without being dismissive. In call analysis, objection handling is often a scored criterion that checks whether concerns were surfaced and resolved.

Discovery questions

Discovery questions are open-ended questions a salesperson asks early in a deal to understand the prospect's situation, needs, pain points, budget, and decision process. They guide the conversation toward whether and how the product fits. Call analysis can detect how many discovery questions were asked and whether key topics were explored.

Next step (call-to-action)

A next step is a specific, agreed action that moves a deal forward after a call, such as a scheduled follow-up meeting, a sent proposal, or an introduction to a decision maker. Securing a concrete next step is a strong signal of progress. Call analysis often checks whether a clear next step was set before the call ended.

MEDDIC

MEDDIC is a B2B sales qualification framework whose letters stand for Metrics, Economic buyer, Decision criteria, Decision process, Identify pain, and Champion. Reps use it to assess whether a deal is qualified and what information is still missing. In call review, MEDDIC fields can serve as criteria to check what was uncovered.

BANT

BANT is a sales lead-qualification framework standing for Budget, Authority, Need, and Timeline. A rep checks whether the prospect has money to spend, the power to decide, a real need, and a timeframe to buy. BANT helps prioritize leads; in call analysis these four areas can be tracked as criteria covered during the conversation.

Win rate

Win rate is the percentage of sales opportunities that result in a closed deal, calculated as deals won divided by total deals pursued over a period. It is a core measure of sales effectiveness. Comparing win rate against call behaviors helps teams see which conversation patterns are associated with closing more deals.

Coaching

Sales coaching

Sales coaching is the ongoing practice of a manager or system helping a salesperson improve specific skills and behaviors, often using real call examples and measurable feedback. Rather than one-off training, it is repeated and targeted. Conversation intelligence supports coaching by pinpointing moments and patterns worth reviewing together.

Call calibration

Call calibration is a session where multiple reviewers score the same call independently and then compare results to align on how the scorecard should be applied. It reduces subjective disagreement and makes scoring consistent across evaluators. Calibration is especially important when both humans and AI score calls against shared criteria.

Ramp time

Ramp time is how long it takes a newly hired salesperson to reach full productivity, such as hitting their expected quota or performance level. It is usually measured in months from start date. Shorter ramp time means faster return on hiring; call review and coaching are common ways teams try to reduce it.

Interview analysis

Structured interview

A structured interview is a hiring interview where every candidate is asked the same predefined questions in the same order and rated on the same scale. This consistency makes candidates more comparable and reduces bias relative to unstructured chats. In MeetGrade, structured interviews are analyzed as decision support based on what candidates say, not facial emotion or lie detection.

STAR method

The STAR method is a framework for answering and evaluating behavioral interview questions, where the response covers Situation, Task, Action, and Result. It prompts candidates to give concrete examples instead of vague claims. In interview analysis, STAR serves as a structure to check whether an answer included real context, the candidate's own actions, and an outcome.

Competency-based interview

A competency-based interview assesses a candidate against specific job-related competencies, such as communication, problem solving, or teamwork, usually by asking for past examples of relevant behavior. Answers are rated against defined competency criteria. As used in MeetGrade, this is evidence-based decision support that evaluates what candidates describe, not personality scoring or emotion reading.